National Academies Press: OpenBook

Applying GPS Data to Understand Travel Behavior, Volume I: Background, Methods, and Tests (2014)

Chapter: Appendix C - Questionnaire Responses from Traffic Data Providers

« Previous: Appendix B - Industry Expert Questionnaires
Page 99
Suggested Citation:"Appendix C - Questionnaire Responses from Traffic Data Providers." Transportation Research Board. 2014. Applying GPS Data to Understand Travel Behavior, Volume I: Background, Methods, and Tests. Washington, DC: The National Academies Press. doi: 10.17226/22370.
×
Page 99
Page 100
Suggested Citation:"Appendix C - Questionnaire Responses from Traffic Data Providers." Transportation Research Board. 2014. Applying GPS Data to Understand Travel Behavior, Volume I: Background, Methods, and Tests. Washington, DC: The National Academies Press. doi: 10.17226/22370.
×
Page 100
Page 101
Suggested Citation:"Appendix C - Questionnaire Responses from Traffic Data Providers." Transportation Research Board. 2014. Applying GPS Data to Understand Travel Behavior, Volume I: Background, Methods, and Tests. Washington, DC: The National Academies Press. doi: 10.17226/22370.
×
Page 101
Page 102
Suggested Citation:"Appendix C - Questionnaire Responses from Traffic Data Providers." Transportation Research Board. 2014. Applying GPS Data to Understand Travel Behavior, Volume I: Background, Methods, and Tests. Washington, DC: The National Academies Press. doi: 10.17226/22370.
×
Page 102
Page 103
Suggested Citation:"Appendix C - Questionnaire Responses from Traffic Data Providers." Transportation Research Board. 2014. Applying GPS Data to Understand Travel Behavior, Volume I: Background, Methods, and Tests. Washington, DC: The National Academies Press. doi: 10.17226/22370.
×
Page 103
Page 104
Suggested Citation:"Appendix C - Questionnaire Responses from Traffic Data Providers." Transportation Research Board. 2014. Applying GPS Data to Understand Travel Behavior, Volume I: Background, Methods, and Tests. Washington, DC: The National Academies Press. doi: 10.17226/22370.
×
Page 104
Page 105
Suggested Citation:"Appendix C - Questionnaire Responses from Traffic Data Providers." Transportation Research Board. 2014. Applying GPS Data to Understand Travel Behavior, Volume I: Background, Methods, and Tests. Washington, DC: The National Academies Press. doi: 10.17226/22370.
×
Page 105
Page 106
Suggested Citation:"Appendix C - Questionnaire Responses from Traffic Data Providers." Transportation Research Board. 2014. Applying GPS Data to Understand Travel Behavior, Volume I: Background, Methods, and Tests. Washington, DC: The National Academies Press. doi: 10.17226/22370.
×
Page 106
Page 107
Suggested Citation:"Appendix C - Questionnaire Responses from Traffic Data Providers." Transportation Research Board. 2014. Applying GPS Data to Understand Travel Behavior, Volume I: Background, Methods, and Tests. Washington, DC: The National Academies Press. doi: 10.17226/22370.
×
Page 107
Page 108
Suggested Citation:"Appendix C - Questionnaire Responses from Traffic Data Providers." Transportation Research Board. 2014. Applying GPS Data to Understand Travel Behavior, Volume I: Background, Methods, and Tests. Washington, DC: The National Academies Press. doi: 10.17226/22370.
×
Page 108
Page 109
Suggested Citation:"Appendix C - Questionnaire Responses from Traffic Data Providers." Transportation Research Board. 2014. Applying GPS Data to Understand Travel Behavior, Volume I: Background, Methods, and Tests. Washington, DC: The National Academies Press. doi: 10.17226/22370.
×
Page 109
Page 110
Suggested Citation:"Appendix C - Questionnaire Responses from Traffic Data Providers." Transportation Research Board. 2014. Applying GPS Data to Understand Travel Behavior, Volume I: Background, Methods, and Tests. Washington, DC: The National Academies Press. doi: 10.17226/22370.
×
Page 110
Page 111
Suggested Citation:"Appendix C - Questionnaire Responses from Traffic Data Providers." Transportation Research Board. 2014. Applying GPS Data to Understand Travel Behavior, Volume I: Background, Methods, and Tests. Washington, DC: The National Academies Press. doi: 10.17226/22370.
×
Page 111
Page 112
Suggested Citation:"Appendix C - Questionnaire Responses from Traffic Data Providers." Transportation Research Board. 2014. Applying GPS Data to Understand Travel Behavior, Volume I: Background, Methods, and Tests. Washington, DC: The National Academies Press. doi: 10.17226/22370.
×
Page 112
Page 113
Suggested Citation:"Appendix C - Questionnaire Responses from Traffic Data Providers." Transportation Research Board. 2014. Applying GPS Data to Understand Travel Behavior, Volume I: Background, Methods, and Tests. Washington, DC: The National Academies Press. doi: 10.17226/22370.
×
Page 113
Page 114
Suggested Citation:"Appendix C - Questionnaire Responses from Traffic Data Providers." Transportation Research Board. 2014. Applying GPS Data to Understand Travel Behavior, Volume I: Background, Methods, and Tests. Washington, DC: The National Academies Press. doi: 10.17226/22370.
×
Page 114
Page 115
Suggested Citation:"Appendix C - Questionnaire Responses from Traffic Data Providers." Transportation Research Board. 2014. Applying GPS Data to Understand Travel Behavior, Volume I: Background, Methods, and Tests. Washington, DC: The National Academies Press. doi: 10.17226/22370.
×
Page 115
Page 116
Suggested Citation:"Appendix C - Questionnaire Responses from Traffic Data Providers." Transportation Research Board. 2014. Applying GPS Data to Understand Travel Behavior, Volume I: Background, Methods, and Tests. Washington, DC: The National Academies Press. doi: 10.17226/22370.
×
Page 116

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

99 AirSage 1. Describe the primary and secondary markets for your data products (i.e., real-time traffic, transportation plan- ning, etc.). AirSage’s primary market is government MPO’s for transportation planning and origin–destination studies. AirSage is also a leader in nationwide traffic flow analysis and an innovative product called FastCache—the next- generation marketing and LBS strategy. 2. List and describe the different data-generating tech- nologies that are used to build your company’s data products (for example, in-dash navigation devices, per- sonal navigation devices, non-GPS cell phones, GPS cell phones, truck GPS/AVL). Is one technology primary; if yes, please identify. AirSage has the exclusive ability to mine signaling data from any type of mobile device from two major carriers (roughly 70% of the U.S. population), and from those location points we can infer traffic speeds, phone/user origins and destinations, and travel trends. 3. If you are using personal mobile devices (such as cell phones or smart phones) as a data source, please describe the penetration rate of this data source. Because we collect data points as a mobile device com- municates with a tower, any type of mobile device is seen. We have 100% penetration rate of mobile devices. 4. Describe your company’s plans for future vehicle or per- sonal technology and data product development. AirSage has spent years getting the technology in place and continues to fine-tune the accuracy of our location data. We built the geolocation systems used by major wireless carriers in North America. We are now com- mercializing the insights and information that can be realized, all while protecting the privacy rights of mobile consumers. AirSage is the leading mobile Big Data pro- vider and can provide clients with solutions in anytime population studies, origin–destination studies, Fast- Cache (geofencing) reporting and mobility attributes. AirSage has partnered with a firm that specializes in demographic and geographic analysis and through this partnership we will be able to garner much more insight into the users’ lifestyle. 5. What is your current geographic coverage? Do you have plans to expand? If so, please elaborate on the planned geography as well as implementation timeline. AirSage has a geographic reach of the entire U.S., includ- ing Hawaii. Wherever a cell phone with one of our two carriers has service, we see it. 6. Describe any of your data products’ limitations that are relevant to road functional classification or other trans- portation system characteristics. Are expected error ranges provided by road segment or TMC location code? For real-time and historical traffic data, AirSage uses TMC codes from FC Class 1-4. Data points are triangulated based on cell signals and assigned a TMC code. Data points are (immediately) tagged as Transient Points (moving) or Activity Points (stationary). 7. Describe how your products are packaged to specifically serve the needs of transportation planners for travel demand forecast modeling or for congestion manage- ment programs. Projects are defined using our standard PO, with the following options: OD Matrix A Trip Distribution table shows the number of people that made trips between a specific Zone pairs during the given time period. A P P E N D I X C Questionnaire Responses from Traffic Data Providers

100 Productions A Productions table shows the number of people that departed from a specific Zone during a specific hour of the day for each hour of each day of the time frame for each of the Zones in the Study Area. Attractions An Attractions table shows the number of people that arrived in a specific Zone during a specific hour of the day for each hour of each day of the time frame for each of the Zones in the Study Area. Traffic Counts A Traffic Count table shows the number of people that traveled a road segment during the given time period(s) for each of the road segments in the Study Area. Home-Work Matrix A Home-Work Matrix shows the number of people who live and work in a specific pair of zones. Sub-Area Analysis A Corridor Study shows the number of people who trav- eled on a specific road segment (corridor) that started and ended their trips in a specific origin, destination, and entered and exited the corridor at specific entry and exit points. Select Link Analysis A Select Link Analysis shows the number of people who traveled a specific road segment (link) while making a trip between specific origin and destination pairs. External Station Analysis An External Station Analysis shows the number of people that traveled through a given metropolitan area, grouped by given count station pairs. Trip Chains Trip Chains show the number of trips within a given geographic area that contain a specific sequence of stops and/or Activity Points for a given time period. Trip Frequency A Trip Frequency table shows the number people who took a specific number or number range of trips from a given origin to a given destination over a specified time period. Trip Duration A Trip Duration table shows the number of people who traveled between the specific Origin and Destination pairs during the specified time frame and stayed for a certain number of days. Anytime Population An Anytime Population table shows the number of people who were present or passed through a given geo- graphic area during a given time period. Traffic Flow A Traffic Flow table shows average travel times for spe- cific road segments during a given time period. This data includes, by individual road segment, traffic flow by day- part at 5-minute intervals. 8. For each of your specific data products (copy and repeat this question and categories as needed), please describe the following: a) Data product name and description b) Raw data frequency and accuracy c) Data cleaning process d) Level of aggregation or disaggregation e) Cost structure Origin–Destination This AirSage suite of analysis and reporting tools lets you analyze billions of location points derived each day from our exclusive network of over 100 million mobile consumer devices. Study traffic zone-to-zone, by trip or travel patterns, even by demographic characteristics. You’ll gain powerful insights that fuel the creation of new businesses, additional locations, or the development of a new product or service. AirSage’s suite of Origin–Destination reports is based on billions of location points derived every day from our exclusive access to over 100 million mobile devices. Data is collected in real time and at lat/long accuracy level but aggregated up to 1,000-m grid cells. Smaller grid cells can be utilized. AirSage’s patent-protected Wireless Signal Extraction (WiSE™) technology aggregates and analyzes signaling data from individual handsets throughout a broad cellu- lar network. In essence, each individual handset becomes a mobile location sensor, allowing AirSage to identify how phones move over time. AirSage validates the informa- tion and converts it into real-time location data, includ- ing traffic speed and time of day. Anywhere our wireless partners provide coverage, AirSage can provide location information in a cost-effective manner, customized for your business. Patented technology and multiple layers of privacy pro- tection ensure that no proprietary, customer-identifying data is accessed or released from within the carrier’s secure environment. We report at the census tract level with at least 10 mobile devices present. All data is bal- anced with time. Cost structure varies from project to project and is based on multiple variables such as population, study area, study length, etc.

101 FastCache FastCache, as a part of an integrated marketing strategy, can drive greater consumer engagement. FastCache is the quickest, most affordable way to get unlimited location fixes for all your mobile opt-in subscribers any time, all the time. You’ll receive continual location updates when- ever a device communicates with the network. AirSage FastCache gives you access to multiple device locations with a single request. It provides you with the ability to take location information, store it, analyze it, and act on the trends and patterns you uncover. Location updates are speedy—with latency of less than a second for single or multiple device locations. Informa- tion is time-stamped and provides a “last seen” location for all your opted-in mobile devices. FastCache works with all mobile devices on any supported network: smartphone or feature phone, with GPS or with- out. No special software, configuration, or user intervention is required. And, unlike GPS-based location services, Fast- Cache won’t drain the batteries on mobile devices. Instant, time-stamped locations wrapped up tight inside a layer of anonymity are sent as often as the client needs them. Patented technology and multiple layers of privacy pro- tection ensure that no proprietary, customer-identifying data is accessed or released from within the carrier’s secure environment. FastCache is available in an easy-to-understand flat rate. Rates begin at one penny per ping, with a max of $.03/per day and/or $0.45/per month, per subscriber. The volume discount of $0.45 per subscriber/per month allows for unlimited “pinging.” Traffic Flow AirSage gives you real-time and historical traffic infor- mation on over 500,000 miles of highway and arterial roads throughout the U.S. With over 100 million mobile devices continually reporting in, AirSage offers up-to- date information for vehicle movement, traffic speeds, patterns and location. Clients get traffic data updated every 1–2 minutes, every day, 24 hours a day. AirSage will deliver detailed, minute- by-minute speed and travel time history for the same nationwide coverage. Clients have access to archived traffic information derived from over 100 million mobile devices, 24 × 7. Testers manually collected GPS data points for 3 major cities and compared them to AirSage’s results. AirSage scored 91%–93% accuracy on real-time conges- tion reports. AirSage’s patent-protected Wireless Signal Extraction (WiSE™) technology aggregates, and analyzes signal- ing data from individual handsets throughout a broad cellular network. In essence, each individual handset becomes a mobile location sensor, allowing AirSage to identify how phones move over time. AirSage validates the information and converts it into real-time location data, including traffic speed and time of day. Anywhere our wireless partners provide coverage, AirSage can pro- vide location information in a cost-effective manner, customized for your business. Patented technology and multiple layers of privacy pro- tection ensure that no proprietary, customer-identifying data is accessed or released from within the carrier’s secure environment. We report at the census tract level with at least 10 mobile devices present. All data is balanced with time. Cost structure varies from project to project and is based on multiple variables such as population, study area, study length, etc. 9. Describe any potential demographic bias that exists in your data sources. According to CTIA US Wireless; “There are now more wireless devices being used in the United States than there are people, and Americans have doubled the amount of Internet data traffic they generate on smart phones, according to the trade group CTIA.” Because we can see any and all mobile devices on our two partner networks, there are no demographic biases. 10. Describe any data usage clauses of agreements that come with new vehicle/device purchases that enable or autho- rize your firm to use personal mobility data. All of our data is collected anonymously behind a firewall; therefore it does not affect the user’s privacy or mobile device capabilities. Through our FastCache product, a person must consent to opt-in for the phones identity to be revealed. 11. Describe how data source privacy and location/time of day details are protected. AirSage spends a great deal of time and effort ensuring the privacy and security of mobile device users. Timestamp information is available and necessary to our studies, but individual location data is always aggregated up to the cen- sus block and census tract level. Unless a phone has opted in, we will never follow an individual phone for a study. AirSage’s privacy module ensures that sensitive personal information is not compromised. It protects consumers from unwanted intrusions, yet enables them to interact with the brands and services they choose. Our technology is fully server-based so there’s no impact to the carrier, the consumer, or your business—no soft-

102 traffic services. No single data source type can provide the accuracy, coverage or scalability that is required in the market today. GPS-based Probe Vehicles and Devices: While INRIX was not the first company to use GPS vehicle probes for traf- fic information, INRIX has built the world’s largest GPS probe network using real-time data from nearly 100 mil- lion probes. These vehicles include cars and commer- cial vehicles such as taxis, limos, airport shuttles, service delivery vehicles, long-haul trucks, and less than truck- load vehicles, plus a rapidly growing number of consumer vehicles. This floating car data is the single biggest input to the INRIX traffic model. Mobile Devices: GPS-enabled smartphones are also becoming an important component of the network. As an example, iPhone users in San Francisco report to INRIX via the INRIX Traffic application as well as other GPS-enabled applications that include INRIX traffic information. Mobile consumers using INRIX Traffic apps on the iPhone and Android-based smartphones have been regularly contributing probe data since the app availability in mid-2009. Road Sensors: More than 30,000 (and growing) road sen- sors across the U.S. Other Flow Sources: Other traffic flow sources including cellular probe data and toll tags. 3. If you are using personal mobile devices (such as cell phones or smart phones) as a data source, please describe the penetration rate of this data source. N/A—This information is proprietary. 4. Describe your company’s plans for future vehicle or per- sonal technology and data product development. INRIX data services include product offerings developed specifically for the public sector, auto OEMs, consumer mobile applications, Internet and media. Detail of future products is proprietary information. 5. What is your current geographic coverage? Do you have plans to expand? If so, please elaborate on the planned geography as well as implementation timeline. From a road coverage standpoint, INRIX provides traf- fic information on over 850,000 miles across the United States with over 600,000 individual TMC links. INRIX currently provides coverage in the U.S., Canada, and Western Europe. In 2013, the company plans to expand to Brazil, Russia, China, South Africa, and Scandinavia, as well as preparing for operations in Eastern Europe, Central America, Australia, India, and northern Africa. ware to load, no stress on systems, and no impact to cell phone battery life. AirSage is regularly tested by indepen- dent security auditors to ensure the data coming in and going out is fully anonymous. In addition, AirSage is compliant with the Telecommu- nications Act of 1996; the Wireless Communications and Public Safety Act of 1999; FCC Proposed Rule-making following the CTIA petition to the FCC on Wireless Location Privacy Principles, November 22, 2000; the European Union Location Privacy, Article 9, amended July 12, 2000; and the individual privacy policies of our carrier partners. INRIX Inc. 1. Describe the primary and secondary markets for your data products (i.e., real-time traffic, transportation planning, etc.). INRIX has a full range of services specifically developed for public transportation agencies. Following is an out- line of available services, aligned with the three compre- hensive strategies typically aimed at fighting congestion. Services for Efficient Operations • Real-Time and Predictive Traffic Flow • Traffic Information Network • Regional Incident Integration Platform • Dynamic Route Travel Times Services for Effective Capacity Planning • Historical Traffic Flow Statistics • Historical Five-Minute Archives • Dynamic Route Travel Time Archive • Analytics and Performance Measures Services to Optimize Demand • White Label Mobile Applications • Custom Mobile Application • Advanced Routing • Consumer Alerting and Planning 2. List and describe the different data-generating technologies that are used to build your company’s data products (for example, in-dash navigation devices, personal navigation devices, non-GPS cell phones, GPS cell phones, truck GPS/ AVL). Is one technology primary; if yes, please identify. INRIX is able to provide extensive traffic information because of the integration of a large variety of sources to calculate the real-time speed on the roadways. INRIX evaluates and uses advanced fusion technologies to intel- ligently integrate the full range of potential traffic data source types in consideration of creating high-quality

103 INRIX Historical Archive is a running archive of real- time speeds provided by INRIX for all TMCs in service at that time. Data provided on each segment for each time slice includes the Speed, Travel Time, and the Score. The archive data is available in 5 minute increments begin- ning July 1, 2008, through December 31, 2010. Starting January 1, 2011, the running archive is available in one minute increments. INRIX data, both real-time and historic, is now avail- able as a subscription service bundled with state-of-the- practice analytics and visualization tools. INRIX has partnered with the University of Maryland to expand the Vehicle Probe Project Analytics Suite currently in use by the I-95 Corridor Coalition to provide national coverage. The Suite provides a real-time dashboard as well as instant access to historical data and the ability to compute common mobility performance measures on demand along with visualization tools. An over- view of the VPP Analytics Suite is available in the form of a short video tutorial at: http://vpp.ritis.org/suite/ screencast/. The analytics suite includes: • A System Dashboard indicating current congestion levels and bottlenecks • Raw data query tool for instant access to archived data based on user specified locations and date ranges • Historical Analytic Tools to instantly calculate com- mon performance measures for user defined corri- dors and date ranges, including: – Average Speed – Travel Time Index – Travel Time – Buffer Index – Buffer Time – Planning Time Index – Planning Time • Visualization Tools for defined performance mea- sures, including: charts, contour plots, and tabular summaries • Bottleneck Ranking Tool to identify system bottle- necks for user defined date ranges 8. For each of your specific data products (copy and repeat this question and categories as needed), please describe the following: a) Data product name and description b) Raw data frequency and accuracy c) Data cleaning process d) Level of aggregation or disaggregation e) Cost structure 6. Describe any of your data products’ limitations that are relevant to road functional classification or other trans- portation system characteristics. Are expected error ranges provided by road segment or TMC location code? N/A 7. Describe how your products are packaged to specifically serve the needs of transportation planners for travel demand forecast modeling or for congestion manage- ment programs. INRIX has several products developed specifically for planners and performance managers. Planners and those interested in model calibration are most interested in aggregation of historical data while performance manag- ers may want to see current conditions in addition to the ability to review historical performance. INRIX provides choices in the type of historical aggregation as well as a new suite of analytic tools and visualizations that can dis- play both real-time and historical performance measures. INRIX Historical Speed Statistics use comprehensive data from the INRIX Smart Driver Network, including billions of historical data points, to provide true histori- cal average speeds and statistical distribution on indi- vidual road segments covering nearly 1 million miles on major freeways, highways, urban and rural arterials and side streets throughout North America. This data is spe- cific to every day of the week, every hour or quarter hour of the day, and is reported at the Traffic Message Channel (TMC) link level or at the smallest road segment for Tele Atlas and NAVTEQ map databases. Using INRIX’s proprietary technology, the data is ana- lyzed and normalized to account for the impact of major events, seasonal traffic patterns, typical weather condi- tions and other variables that can impact traffic flow— ensuring the highest degree of accuracy. The data is updated regularly, incorporating both changes to map databases as well as additional historical data from the INRIX Smart Driver Network. INRIX statistical parameters provided for each time bin include the average speed, standard deviation, the 5th, 10th, 15th, 20th, 25th, 30th, 40th, 50th, 60th, 70th, 75th, 80th, 85th, 90th, and 95th percentiles of speed values in mph, and the 30, 40, 50, 60, 70, 80, and 90 mph “failure rates,” or the percentage of data points that fall below the specific speed threshold for the given segment/time. In addition to typical annual files, monthly and quarterly files are also available. Monthly and quarterly files are available starting January, 2009. Future monthly files for 2012 and beyond can be made available by the 15th of the following month.

104 INRIX Historical Archive INRIX Historical Archive is a running archive of Real- Time speeds provided by INRIX for all TMCs in service at that time. Data provided on each segment for each time slice includes the Speed, Travel Time, and the Score. The archive data is available in 5-minute increments beginning July 1, 2008, through December 31, 2010. Starting January 1, 2011, the running archive is available in one minute increments. INRIX Real-Time Traffic Information Network INRIX’s Real-Time Traffic Information Network pro- vides a visualization of real-time traffic information via a hosted statewide traffic monitoring website of the real- time coverage so agencies can monitor conditions on roadways in real time at Traffic Operations Centers and for other internal agency viewing/use. The monitoring website includes access for 20 simultaneous users (more users are available for additional amount depending on number of users). INRIX Traffic Tile Overlays Traffic Tile Overlays allow for a web services API request to INRIX and, in response, agencies receive a semi- transparent, match projection (Mercator) overlay image of INRIX Traffic Flow for display on common mapping platforms (Google, Microsoft, Yahoo!, etc.). INRIX Traffic Tile Overlays are fully configurable and are available in various tile sizes (256 x 256 standard) with broad browser support (IE, Firefox, WebKit [Safari, Chrome]) and token-based security (access via token with timeout, configurable to a user). INRIX Dynamic Route Travel Times INRIX Dynamic Route Travel Times provide the current travel time (based on real-time traffic conditions) along a precisely-specified route between any Origin (starting point) and Destination (ending point) in either direction. Travel times can be queried in real time with archive rights to provide information for both instantaneous and histor- ical analytic purposes This information may be used for: corridor management; planning and modeling purposes; reliability and performance measures; making opera- tional decisions; and disseminating traveler information like travel times on dynamic message signs (DMS). INRIX Analytical Tools INRIX data, both real-time and historic, is now available as a subscription service bundled with state-of-the-practice analytics and visualization tools. See full description and figures in response to Question 7. 9. Describe any potential demographic bias that exists in your data sources. N/A Following is an overview of specific INRIX Data Services specifically developed for public-sector transportation agencies. Cost of services costs are based on annual terms but can be customized. Cost structure is proprietary as most agencies contract for specific services through a competitive process. INRIX Real-Time Flow Data Service Real-Time Flow is INRIX’s full suite of traffic data which is available via an API call as often as once per minute. The data provided via XML includes road segment code; roadway name and cross streets of roadway; time; current speed in mph (“speed”); typical speed in mph (“aver- age”); free flow speed in mph (“reference”); and travel time along segment in minutes (“traveltimeminutes”). Predicted speeds and travel times are also available via an API call. Predicted times include 15 minute, 30 minute, 60 minute, 24 hours, and 48 hours into the future at a minimum. INRIX Historical Traffic Flow Statistics INRIX Historical Traffic Flow Statistics use comprehen- sive data from the INRIX Smart Driver Network, includ- ing billions of historical data points, to provide true historical average speeds and statistical distribution on individual road segments covering over nine thousand centerline miles on major freeways, highways, urban and rural arterials and side streets throughout Washington State. This data is specific to every day of the week, every hour or quarter hour of the day, and is reported at the Traffic Message Channel (TMC) link level. Using INRIX’s proprietary technology, the data is ana- lyzed and normalized to account for the impact of major events, seasonal traffic patterns, typical weather condi- tions and other variables that can impact traffic flow— ensuring the highest degree of accuracy. The data is updated regularly, incorporating both changes to map databases as well as additional historical data from the INRIX Smart Driver Network. INRIX statistical parameters provided for each time bin include the average speed, standard deviation, the 5th, 10th, 15th, 20th, 25th, 30th, 40th, 50th, 60th, 70th, 75th, 80th, 85th, 90th, and 95th percentiles of speed values in mph, and the 30, 40, 50, 60, 70, 80, and 90 mph “failure rates,” or the percentage of data points that fall below the specific speed threshold for the given segment/time. In addition to typical annual files, monthly and quarterly files are also available. Monthly and quarterly files are available starting January, 2009. Future monthly files for 2012 and beyond can be made available by the 15th of the following month.

105 applications. Many of the clients who use our navigation applications provide data back to Nokia. The mix of data, coupled with fixed point traffic monitoring devices, cre- ates extensive coverage on freeway and arterial roadways using sophisticated algorithms to produce reliable data for Nokia’s private and public clients. 3. If you are using personal mobile devices (such as cell phones or smart phones) as a data source, please describe the penetration rate of this data source. Nokia receives billions of GPS data points monthly globally from cell phones and other smart devices. The penetration rate varies by geographic area, time of day and day of week. 4. Describe your company’s plans for future vehicle or per- sonal technology and data product development. Nokia is constantly working to expand and enhance the data, coverage and applications. A few examples of our latest products include Nokia Drive, Nokia Transit, and Mirrorlink. Nokia Drive and Nokia Transit applications allow users to interact more naturally with the world sur- rounding them and helps travelers determine their best routing options. Mirrorlink enables the phone to interact with in-vehicle devices integrating the smartphone into the dashboard to provide safe and seamless access within the connected vehicle. 5. What is your current geographic coverage? Do you have plans to expand? If so, please elaborate on the planned geography as well as implementation timeline. In the United States the Nokia coverage includes all the Interstate highways and limited access roadways, in addi- tion to thousands of miles of arterial roadways in metro- politan areas. We continue to expand the arterial footprint to more rural areas and will soon be expanding to include additional major arterial roadways nationwide. 6. Describe any of your data products’ limitations that are relevant to road functional classification or other trans- portation system characteristics. Are expected error ranges provided by road segment or TMC location code? The roadways with higher volume and higher functional class roadways also carry more available probe vehicles. Lower functional classification roadways have many com- plex movements and require more sophisticated algo- rithms to match data and determine the speeds associated with a smaller arterial. To provide users with a mechanism to manage data that is distributed in time and space and calculated for each roadway segment, Nokia provides a confidence level in its real-time traffic data feed. The con- fidence value provides an indication of the quality and 10. Describe any data usage clauses of agreements that come with new vehicle/device purchases that enable or autho- rize your firm to use personal mobility data. N/A 11. Describe how data source privacy and location/time of day details are protected. INRIX assures complete privacy by requiring all data providers remove all personally identifiable information from all data prior to sending to INRIX. Nokia/NAVTEQ 1. Describe the primary and secondary markets for your data products (i.e., real-time traffic, transportation plan- ning, etc.) Nokia is successfully delivering traffic data to government agencies and private sector companies globally. Nokia’s primary market is real-time data for traveler information and secondary markets are traffic operations and trans- portation planning, including archived data. Nokia North American clients include over 20 State Departments of Transportation, Verizon Wireless, Sirius XM Satellite Radio, Microsoft Bing, Garmin, The Weather Channel, and Comcast. We also provide traffic data services and applications to all of the major United States mobile phone carriers: AT&T, Sprint, T-Mobile, and Verizon. Collectively, more than 220 million people in North America are served by Nokia traffic. Nokia Traffic powers: • 20 out of the 20 top ranked in-vehicle navigation sys- tems, supporting over 150 vehicle models; • 75% of traffic enabled personal navigation devices (PNDs), including 3.1 million PND users with Ad-supported free lifetime traffic; and 3.5 million paying subscribers • Major online mapping applications including Bing Maps • Wireless providers including Nokia, Verizon’s VZ Navigator, RIM/Blackberry, and Gokivo the first traffic enabled Turn-By-Turn app for the iPhone 2. List and describe the different data-generating technologies that are used to build your company’s data products (for example, in-dash navigation devices, personal navigation devices, non-GPS cell phones, GPS cell phones, truck GPS/ AVL). Is one technology primary; if yes, please identify. Nokia’s traffic is powered by a variety of originating probe devices including GPS-enabled personal navigation devices, commercial fleets and commercial third-party

106 updates are provided every two minutes. Accuracy for an individual TMC depends on the number of data sources the age of the data and the variability of traffic conditions. The data is quality checked prior to data integration, map matched in real time, cleaned for erroneous values, and filtered extensively. The cleaning and matching process is more extensive on arterial roadways due to increased variability of travel patterns requiring more sophisticated map-matching and algorithm processing to support quality results. The data is collected from consumer and commercial probes, toll tags and sensors. The average speed and travel time is aggregated for each TMC using a sophisticated algorithm. Nokia’s diverse user community of public and private sec- tor clients demands a highly flexible and adaptable licens- ing program. Several licensing options are available to ensure scalable and effective use of Nokia data. Licensing fees are determined based on the number of users, licens- ing term (i.e., number of years), geographic extent, and data delivery mechanism (e.g., desktop, web based). Fees are not based on the miles of roads covered in a region. Archived Data Nokia maintains several types of historic traffic data that is summarized in the Nokia Traffic Patterns and Traffic Analytics products. Traffic Patterns applies a weighting algorithm to multiple years of traffic data to provide the best possible assignment of the typical speeds experienced on all roads and highways. Traffic Analytics is an annual summary of traffic speeds contain 15-minute intervals by day, month and season. Both models also represent holiday travel conditions. Confidence scores are provided with all data summaries. Nokia also summarizes data directionally at the link level and TMC level for all roads. Raw GPS data is used to develop the archived products. During the data cleaning process for the data archive, erroneous values are removed and a raw probe is refor- matted to enable map matching. The output of our clean process normalizes the data and ensures uniformity and quality prior to data aggregation. The data cleaning occurs on data obtained from the archive of the real-time traffic services as well as from probe archives received directly from third-party vendors. The data is collected from individual probe points, then it is aggregated into 15-minute speeds and available at the TMC level and detailed Nokia link level for all mapped roads. density of the data in time and space. There are also spe- cific challenges in very complex urban conditions includ- ing tunnels and bridges structures with multiple levels that we are continually working to make as precise as possible. 7. Describe how your products are packaged to specifically serve the needs of transportation planners for travel demand forecast modeling or for congestion manage- ment programs. Nokia has an application called Traffic Patterns that includes archived data in a variety of aggregations. The most commonly used format is 15-minute average speed and travel time for each TMC at 15-minute increments for each day of the week. Nokia currently provides real-time and archived data to Michigan DOT. This data will be integrated into the RITIS software application provided by the University of Mary- land for performance measure and other transportation planning. 8. For each of your specific data products (copy and repeat this question and categories as needed), please describe the following: a) Data product name and description b) Raw data frequency and accuracy c) Data cleaning process d) Level of aggregation or disaggregation e) Cost structure Real-Time Data Feed Nokia Traffic Satellite—Nokia is the exclusive provider of real-time traffic services for all satellite radio providers who offer traffic information via satellite radio in North America. Nokia Traffic RDS—Real-time traffic delivered over FM radio using a radio data system (RDS) sub-carrier chan- nel. RDS is well suited for auto companies and PND manufacturers. Nokia Traffic ML—Real-time traffic designed for mobile and server-based navigation, as well as mapping applications. Nokia Traffic Online—Real-time traffic available via consumer traffic websites. Nokia Traffic Digital—Real-time traffic delivered over digital radio’s high bandwidth will mark a major leap for- ward as additional data services beyond traffic become available. Real-time data is continuously collected, models are updated every minute and the files containing the

107 • Before and after studies—changes to road infra- structure, construction • Performance analysis of intersections • Traffic model calibration Geo-Marketing • Site location • Advertising display location • Demographic travel patterns Logistics • Fleet management (See also Navigation below) • Supply chain optimization (See also Navigation below) • Delivery scheduling (See also Navigation below) Navigation (Automotive, Personal Navigation, Internet, Mobile) • Estimated Arrival Time calculation based on day of week, time of day • Time specific route selection: Routing based on day of week, time of day Insurance • Risk assessment—accident hotspot and high risk area identification Live Traffic Transportation Planning/Traffic Engineering: Private/ Public-Government • Active traffic management • Traffic Monitoring—Traffic Control Centers • Flow Data—Speed, Travel Time • Incident Data—Traffic Jam, accident, closure etc. • Traffic website—511 etc. • Variable message sign display—travel time, delay time Navigation • Dynamic live navigation/routing • Dynamic pre-trip route planning • Dynamic Estimated Time of Arrival calculation 2. List and describe the different data-generating techno l- ogies that are used to build your company’s data prod- ucts (for example, in-dash navigation devices, personal navigation devices, non-GPS cell phones, GPS cell phones, truck GPS/AVL). Is one technology primary; if yes, please identify. TomTom data sources include connected (GSM enabled) and non-connected after-market GPS devices, in-dash GPS systems, commercial vehicle GPS systems, GPS smart phones, and third-party incident data. After-market GPS devices represent TomTom’s primary data source. Nokia’s diverse user community of public and private sector clients demands a highly flexible and adaptable licensing program. Several licensing options are avail- able to ensure scalable and effective use of Nokia data. Licensing fees are determined based on the number of users, licensing term (i.e., number of years), geographic extent, and data delivery mechanism (e.g., desktop, web based). Fees are not based on the miles of roads covered in a region. 9. Describe any potential demographic bias that exists in your data sources. We receive data from a variety of data source types that likely mitigates bias in the data sources. Approximately half of the probe data is consumer, approximately half is commercial. The consumer mix includes mobile phones and mobile phone applications, personal navigation devices and navigation systems. 10. Describe any data usage clauses of agreements that come with new vehicle/device purchases that enable or autho- rize your firm to use personal mobility data. Many of Nokia’s clients provide data back to Nokia and therefore provide the data usage clause to the user. Some of Nokia’s products simply ask “(Product name) would like to use your current location, OK?” 11. Describe how data source privacy and location/time of day details are protected. All data collected has a unique ID and there is no way to track the information back to an individual user. As an example, Nokia uses a completely anonymous id that resets periodically to further ensure anonymity. TomTom 1. Describe the primary and secondary markets for your data products (i.e., real-time traffic, transportation plan- ning, etc.) TomTom markets served include consumer navigation systems, commercial vehicle navigation systems, in-dash solutions for automotive OEMs, traffic information sys- tems (e.g., 511), transportation planning and modeling, and GIS analysis. Historic Traffic Transportation Planning/Traffic Engineering: Private/ Public-Government • Network performance monitoring • Road network bottleneck reporting/analysis • Noise and emission hotspot identification

108 crowd-sourced GPS speed measurements from devices and smartphone applications as well as third-party information on road closures and accidents. Industry standards for data transmission are used and the traffic information can be provided to public and private enti- ties using standard protocols. Individual privacy is of paramount importance in TomTom’s systems. Because a large proportion of real- time information is crowd-sourced, the crowd has to be able to trust that their information will not be misused or shared inappropriately. Protecting privacy goes beyond legal limitations: if the crowd does not perceive us as trustworthy, the crowd will no longer be a source. Tom- Tom has developed methods of safeguarding the privacy of individuals and the information they provide. These continue to evolve as the number of channels providing crowd-sourced information grows. 5. What is your current geographic coverage? Do you have plans to expand? If so, please elaborate on the planned geography as well as implementation timeline. TomTom provides historical traffic data for 45 countries, and real-time traffic data in 23 countries. Coverage is expanding to include China and Russia this year and will continue to expand in the future. 6. Describe any of your data products’ limitations that are relevant to road functional classification or other transportation system characteristics. Are expected error ranges provided by road segment or TMC loca- tion code? Limitations: probe-based speed measurements rely on there being vehicles on the road to provide data. Small, local roads with little or no traffic have therefore the lowest coverage in our system, yet we do provide speeds based on multiple years of measured, historical speed information by time of day and day of week. Some of our products include a ‘confidence value’ for each road sec- tion to indicate the number of recent observations taken into account with that updated value in the file. Due to limitations of TMC table coverage, TomTom has imple- mented new location referencing which can provide traf- fic information at any location (see OpenLR.org). 7. Describe how your products are packaged to specifically serve the needs of transportation planners for travel demand forecast modeling or for congestion manage- ment programs. TomTom Traffic Stats TomTom lets you access the largest historic traffic database in the world available for governments and enterprises via our web-based Traffic Stats portal. Measure location 3. If you are using personal mobile devices (such as cell phones or smart phones) as a data source, please describe the penetration rate of this data source. The specific penetration rate is unknown. There have been in excess of 1 million downloads of the TomTom smartphone GPS navigation application globally—but TomTom only collect traffic traces from devices docked in a car holder as a better indication that they are being used on a vehicle journey as opposed to train/bus etc. TomTom also uses cellular probe data as an input source. By looking at the activity of cell phones moving near GSM network antennae, the (anonymous) handset loca- tion can be matched to the road network and speed information calculated. Over 80 million GSM probes contribute to the TomTom real-time traffic system in Belgium, France, Germany, Italy, Netherlands, Portugal, Switzerland, and the UK—but always as a supplement to GPS probe data not as a substitute for GPS data. 4. Describe your company’s plans for future vehicle or per- sonal technology and data product development. TomTom will continue to focus on smartphone appli- cations, in-dash systems, and after-market navigation devices to ensure that the richest set of accurate GPS real- time traces from a broad range of vehicles are available for creating traffic information. There will also be a continued focus on traffic content, quality, coverage and geo-expansion to make traffic infor- mation available in more markets globally. TomTom will add increased probe quantity both in real time and for the historical traffic database by adding 3rd party partner GPS data and connected device information starting from sec- ond half of 2012 to improve the accuracy and confidence in the data. The project to be undertaken which will facili- tate addition of the connected, (live,) and 3rd party probes to the historical database will not only increase sample size but also improve the freshness of the data available for analysis. This will take place across all markets where TomTom operates including North America. TomTom has technology for multiple platforms. These include after-market consumer navigation devices, smart- phone applications, fleet management systems and in-dash solutions. Predictive (based on historical information) and real-time route information is also made available on TomTom’s web platform. All our systems are supported by a customer care division, which has won a series of JD Power awards in recent years. TomTom’s back-office systems include our own data fusion technology and provide updated traffic infor- mation every minute. The data fusion uses anonymous

109 may also be possible though it was created to specifically improve route and ETA calculation. This product links directly to the TomTom MultiNet map database. This product is based on a two year average of speed per time of day, (5-minute time bins,) and is released quarterly. 8. For each of your specific data products (copy and repeat this question and categories as needed), please describe the following: a) Data product name and description: TomTom Custom Travel Times Custom Travel Times is a traffic solution designed to give government agencies and enterprises more insight into traffic flows on a specific roadway, or series of connected roadways, (created using an A to B route calculation.) Custom Travel Times provides highly granular speed and bottleneck information for roads around the world. TomTom’s ever-expanding historical traffic database has over 5 trillion data points with over 6 billion new records being added each day—with some roads having more than 20,000 measurements. This makes it possible to obtain actual driven travel times and speeds on any stretch of road over any period of time and time of day. Cus- tom Travel Times covers all roads, from major highways to local and destination roads, throughout Europe and North America. Please find detailed data output informa- tion below. • Excel – Route Name: Customer defined – Time collection: Customer defined – Length (meters)—Total length of the route. – Sample size: Average per segment. – Average Travel Time (hh:mm:ss)—Arithmetic average of travel time over the route. – Median Travel Time: (hh:mm:ss) – Average speed: (kph/mph)—Harmonic average speed. – Travel Time ratio: Comparison Sets – Percentile travel times: 5%–95% • KML – Average speed: KPH/MPH – Length of segment: Meters – Average travel time: Seconds – Median travel time: Seconds – Standard deviation: Seconds – Sample size – Travel Time Ratio: Comparison Sets • Charts – Available online for viewing. Not downloadable. – Charts/Graphs can be made from the data con- tained in the Excel output. accessibility for site selection, identify road network bot- tlenecks and noise emission hotspots, perform before and after studies relating to infrastructure changes or analyze intersection design and performance. All you need is an Internet connected computer, and you will receive 24 × 7 access to TomTom traffic data. • Access to TomTom historical traffic products any- where & anytime. • Tailor-made reports available within 24 hours. • Data can be downloaded for use in other applica- tions/traffic modeling tools, etc. Through the Traffic Stats website three products, Custom Area Analysis, Custom Travel Times and Custom Probe Counts can be accessed. Below please find brief descrip- tions of these three products, with a more detailed descrip- tion to follow in response to question 8. NOTE: Custom Probe Counts has yet to be officially released. Custom Area Analysis Custom Area Analysis delivers a Shape file (*.shp) with both the road geometry of the segments analyzed and the data base, (*.dbf) with the related statistical data for each road segment. This data can be readily uploaded into standard GIS tools to visualize and manipulate the data. A sample size is provided upon completion of the query and one can reject a report before accepting if the sample size proves inadequate for a given project. This allows for the time parameters to be adjusted so that a greater sample size might be obtained. NOTE: This is also the case for Custom Travel Times which is outlined immediately below. Custom Travel Times Results of a Custom Travel Times query are given in three different types of output: • Excel • KML • Charts NOTE: Charts are only available for viewing within the Traffic Stats web portal. However, graphs can be made using the Excel output. Custom Probe Counts The results are given in an industry standard ESRI Shapefile compatible dBASE file and can be downloaded through a simple web interface, once confirmation has been given that the job has been processed. Speed Profiles A fourth historical speed product is also available for the purpose of providing real-world data for the calcula- tion of routes and estimated times of arrival in routing/ navigation applications. Other potential uses of the data

110 by aggregating billions of GPS measurements to offer precise speeds for specific times of day and days of the week. With Speed Profiles routes adapt dynamically to the time of departure and incorporate local knowledge. The optimal route on Monday morning may differ on Sunday afternoon, just as the travel time on a Tues- day in December will be longer than in May. Armed with Speed Profiles, ETAs are highly accurate and travel time is reduced along with stress levels, travel costs and environmental impact. Deployment effort is minimal thanks to a compact data footprint and with wide coverage of highways, urban, rural and second- ary roads it delivers a seamless country by country experience. TomTom Enterprise Traffic TomTom Enterprise Traffic provides precise locations and delays caused by congestion on the road network, allowing routing programs to provide the fastest route based on actual current travel times. By incorporating TomTom Enterprise Traffic into a navigation solution, drivers can determine the quickest route to their desti- nations by considering “live” road conditions. The data in each Enterprise Traffic file includes road delays allow- ing routing programs to evaluate the true travel time to each destination. This product can also be deployed for display purposes such as in traffic control center or even embedded into a website for A to B dynamic rout- ing. The Enterprise Traffic XML feed is made available for consumption via the client pull method every minute in standard formatting and can be implemented using either TMC segments or OpenLR. TomTom HD Flow TomTom HD Flow delivers a real-time, detailed view of traffic speeds on the entire road network, designed for easy integration into traffic management systems or cal- culating current routing travel times. The data output is generated from TomTom’s proprietary fusion engine and is refreshed every minute. For each road segment this delivers the road’s identification, the total travel time under current and ideal conditions and the average speed and quality for that segment. This enables traffic control centers to determine relative levels of road service over wide areas, and personal navigation device and smart- phone manufacturers to benefit from dynamic routing and display. With HD Flow we help you in the placement of real-time traffic signs showing the fastest road choice to a common destination. A quality/confidence value is also provided in the output. The data is made available in XML feed, client pull method, every minute, DATEX2 format, TMC application. TomTom Custom Area Analysis Custom Area Analysis is a traffic solution designed to give government agencies and enterprises more insight into traffic flows for complete road networks. Custom Area Analysis provides highly granular speed and bottle- neck information for large and small road networks around the world across all road classes. TomTom’s ever- expanding historical traffic database has over 5 trillion data points with over 6 billion new records being added each day—with some roads having more than 20,000 measurements. This makes it possible to obtain actual driven travel times and speeds on any stretch of road over any period of time and time of day. Custom Area Analysis covers all roads, from major highways to local and destination roads, throughout Europe and North Amer- ica. No need to purchase road collection hardware such as cameras, and loop sensors. With Custom Area Analysis it is possible to gather statistics on the road network as a whole, from Interstates through arterials to the local street level. • Shapefile (*.dbf) output fields – ID: Segment ID number – Average Travel Time: Harmonic average travel time. kph or mph – Median Travel Time: Seconds – Average Speed: Arithmetic Average. kph—mph – Median Speed: kph or mph – Standard Deviation of Speed: kph or mph – Sample Size: Per segment – Travel Time ratio: Comparison Sets. Seconds – Percentile travel times: 5%–95% Custom Probe Counts Not all studies of road usage require speed and congestion information. Sometimes it is interesting to understand the relative traffic volumes that use the roads around sites being considered for locating new buildings or shopping areas—or even which locations might attract more viewers for an advertising campaign. Custom Probe Counts can provide data on the number of devices that were recorded on individual MultiNet road segments for any given period back to 2008. Highly granular data from TomTom’s ever- expanding historical traffic database has over 5 trillion data points with over 6 billion new records being added each day—with some roads having more than 20,000 mea- surements. Using this data, analysts can quickly gather an indication of the relative traffic volume on the individual roads surrounding the location of interest. Speed Profiles Speed Profiles is different. It is a comprehensive data- base of actual historic roadway speeds. These are attained

111 Real-Time Output Quality: • TUV Certified • Regular QKZ tested (Enterprise Traffic) • HD Flow tested by methodology developed by the Texas Transportation Institute c) Data cleaning process: All data inputs are processed through the proprietary TomTom applications to remove possible errors such as map-matching anomalies, outliers etc. Filters are applied to remove possible data from devices used on public transport/pedestrians and also where vehicles are tem- porarily in gas stations etc. d) Level of aggregation or disaggregation: With respect to aggregation at the road element level TomTom offers different levels of aggregation. In Cus- tom Area Analysis and Custom Travel Times road ele- ments can be quite short, (multiple road elements within one city block for example,) to road elements on high- ways which can be measured in terms of miles. In Live Traffic offerings Traffic Messaging Channel links can be utilized which result in much longer road elements and cover only the TMC network. Real-time data can also be made available in the above mentioned shorter road ele- ment lengths via our open-source OpenLR format which allows for the application of traffic information across the road network wherever information is available. With respect to Live Traffic updates are made available via the client pull method every minute. The historical products Custom Travel Times and Custom Area Analysis data can be sliced by one hour combinable time bins. Future iterations of these two products will offer more flexibility with respect to time bins the details of which are not available at this time. Speed profile data represents two year averaged data which is applied to the TomTom MultiNet map at the road element level. e) Cost structure Custom Travel Times: Price per route. Custom Area Analysis: Price per geographic scope, road class and number of queries. Custom Probe Counts: Price per mile. Real-Time Traffic: Price per mile. 9. Describe any potential demographic bias that exists in your data sources. When TomTom first entered the mass market for navi- gation we saw a bias toward men between the ages of 25 and 35 who drive often. We now know from our own market research that our customer base is now much TomTom HD Route Times TomTom HD Route Times is a turnkey solution pro- viding highly accurate real-time travel and delay times for a specific route either on a temporary basis or for permanent solutions. Key to this data is its flexibility. Travel times and delay statistics are delivered without the need to build infrastructure or install and maintain hardware. Data can be sent continuously to roadside information systems or on a temporary basis to mobile systems used during road works. The data is refreshed every minute allowing traffic control centers to deploy variable message signs (VMS) suggesting alternative routes. Event managers can schedule travel information to be displayed around special events. Even kiosks and corporate offices can take advantage of TomTom HD Route Times’ detailed, granular data for specific local route data. b) Raw data frequency and accuracy: TomTom GPS data In raw form, TomTom GPS data has the following characteristics: • Time stamp: year, month, day, hour, minutes, seconds, hundredths of seconds • Position stamp: latitude and longitude • Identification number (randomly generated) Data from real-time GPS devices are transmitted between every 40 and 60 seconds to a server after which it is pro- cessed through the proprietary TomTom fusion engine. For privacy protection the identification number is changed for every 24-hour period. Further processing makes it possible to calculate: • Speed • Travel time • Speed and travel time variance • Acceleration rate • Deceleration rate • Trip origin, destination, route choice To calculate these statistics, the GPS data points must first be ‘map-matched’. This is a process developed by TomTom to match each GPS point to a specific road seg- ment as defined in our own map database. The map- matching process also performs filtering of unreliable GPS measurements. Data presentation: • Data made available every minute • XML Feed • Datex2 format • Location referenced by TMC code or OpenLR

112 unique serial numbers, potentially allowing (re-)iden- tification of the user are destroyed either immediately or within 20 minutes after their device or car has been shut down. Users also are informed about the fact that TomTom uses their information in an anonymous way to enhance its products and services, which also are made available to business and governments. In those cases geolocation information is obtained from third-party sources, TomTom ensures, technically, orga- nizationally and contractually, the data it receives from the third party does not allow TomTom to identify or even single out the individual contributing the data: TomTom obtains anonymous data only and uses it only for the agreed purpose after which the data is destroyed. TomTom applies advanced Privacy Enhancing Techno l- ogies and organizational measures to subsequently live up to the agreement with its users and third parties. All geolocation data is protected against unauthorized access (i.e., anyone except TomTom) with strong encryption while stored on end user devices and while in transit. To avoid identification or singling out individuals, TomTom irreversibly destroys any unique identifiers immediately upon reception of the data from its users. It those cases where this is not possible (specifically: generating traffic information), one-way pseudonyms with a short lifetime are used and data is kept in volatile memory to not cre- ate potentially recoverable copies of identifiable geoloca- tion data. Pseudonym lifetime is capped at a maximum of 20 minutes after the device contributing the data has been switched off or 24 hours, whichever is shorter. In those cases where TomTom retains a copy of the geolo- cation data, this always is done without any identifying elements (such as device unique serial numbers) and on a per trip basis only, i.e., without maintaining the relation- ship between trip originating from the same device. Please also refer to www.tomtom.com/yourdata for our publicly available information regarding the way Tom- Tom treats information obtained from its customers. TrafficCast 1. Describe the primary and secondary markets for your data products (i.e., real-time traffic, transportation plan- ning, etc.) Primary markets: • Provide real-time traffic data (Dynaflow and inci- dents) to support online and mobile applications and to TV stations for broadcasting. • Provide BlueTOAD services to public agencies for workzone impact study, special events traffic flow more representative of the general population, but we still have some underrepresentation of women and of people older than 65 years of age or short-distance com- muters. There may also be some bias against very low household income groups. 10. Describe any data usage clauses of agreements that come with new vehicle/device purchases that enable or autho- rize your firm to use personal mobility data. After-market devices include an opt-in question. In-dash systems vary. Opt-in depends on the actual OEM and the specific agreement. 11. Describe how data source privacy and location/time of day details are protected. Individual privacy is of paramount importance in TomTom’s systems. Because a large proportion of real- time information is crowd-sourced, the crowd has to be able to trust that their information will not be misused or shared inappropriately. Protecting privacy goes beyond legal limitations: if the crowd does not perceive us as trust- worthy, the crowd will no longer be a source. TomTom has developed methods of safeguarding the privacy of individ- uals and the information they provide. These continue to evolve as the number of channels providing crowd-sourced information grows. TomTom uses various data sources to create its maps and map related products and services, such as real-time and historic traffic information. These data sources include geolocation data obtained from indi- viduals, who need to be able to trust TomTom to use their data in a responsible way that does not violate their privacy. To obtain the highest possible yield, resulting in the high- est possible quality, it is paramount for TomTom to foster this trust. As such TomTom is committed to acting above and beyond the OECD privacy principles (notice, purpose, consent, security, disclosure, access and accountability), also laid down in legislation and enforced by independent regulators in the various countries in which TomTom operates. More specifically, in all cases where TomTom obtains geolocation data from its customers, this is done based on prior, explicit, informed consent, which can be with- drawn at any time. Effectively this means that users of TomTom products are informed about the use of their geolocation information and voluntarily opt-in, before any geolocation information is captured to be sent to TomTom for further use. Users are informed about the data being captured and the fact that it will only be used for map, road, traffic and traffic related purposes, under the moniker “we profile roads, not people.” Users also are informed that TomTom will use the data anonymously, i.e., elements, such as account names, email addresses or

113 the process of investigating its critical attributes required for generating useful traffic information. The prelimi- nary findings indicate the needs of adopting theory developed in the linear dynamic system to make this data useful. The final goal of this exercise is to come out a soft- ware product that runs on either device side or server side, which ever mobile data is available, so the traffic information can be produced on the fly to feed mobile application needs. Another product plan that is directly related to end users and has already been in the testing phase is the traffic broadcaster application that is currently under reviewed by media companies. This application is designed to pro- duce graphical and narrative traffic information for live TV and radio broadcast. The requirement of this initia- tive also calls for providing mobile traffic application for smartphone so media can use it as name branding and promotion vehicle. In return, the data collected by these applications will be used to enhance TrafficCast’s product offering. As for the BlueTOAD, TCI has developed several related products to meet both real-time traffic management and offline transportation study needs. Please see http:// trafficcast.com/products/view/blue-toad/ for the details. 5. What is your current geographic coverage? Do you have plans to expand? If so, please elaborate on the planned geography as well as implementation timeline. Currently TCI Dynaflow and incident data service cover- age for functional class 1 to 4 roads in the U.S. BlueTOAD deployment covers major road and arterial in 20+ states in the U.S. as well as in Vancouver, Calgary, and Kelowna, Canada, Sao Paulo, Brazil, Santiago and Puerto Montt, Chile, and Hong Kong, China. The planned geographic expansion includes providing top 20 markets of Dynaflow in Canada by 2013 Q1 and continuously expands BlueTOAD deployment to major metropolitan areas in the U.S. and Central and South America in 2013 and 2014. 6. Describe any of your data products’ limitations that are relevant to road functional classification or other trans- portation system characteristics. Are expected error ranges provided by road segment or TMC location code? Dynaflow: It covers functional class 1 to 4 roads in the U.S. The error rate varies from 10% to 35% depending on the markets and road functional class. Freeway and major arterial in an urban area usually have a higher accuracy than that of minor roads. This is probably due to flow disruption caused by the deployment of monitoring, OD study, and real-time travel speed and travel time reporting. Secondary markets: • Provide real-time and historical speed data to public agencies for transportation planning and modeling. 2. List and describe the different data-generating techno l- ogies that are used to build your company’s data products (for example, in-dash navigation devices, personal navi- gation devices, non-GPS cell phones, GPS cell phones, truck GPS/AVL). Is one technology primary; if yes, please identify. For Dynaflow product, major input data come from two main sources: • GPS data from fleet and mobile device that TCI pur- chases and data exchanges from fleet management companies and business partners that provide location-based service through application installed on GPS-enabled mobile phone. • Data collected by BlueTOAD installed across more than 20 states in the U.S.A. and in other countries such as Canada, Chile, and Brazil. See http://trafficcast. com/products/view/blue-toad/ for the detailed infor- mation about BlueTOAD. • The majority of mobile data used to produce Dyna- flow come from GPS data collected through fleets. For incident data service: • More than 120 Java programs are developed to retrieve and parse publically available incident data from city, state DOTs and 511 systems. • Data collected by TCI operators at National Opera- tion Center based in Madison, Wisconsin, and Phila- delphia through TrafficCaster, an interactive map and table driven web-based application developed by TrafficCast. Operators watch the traffic camera and lis- ten to police radio scanner to identify and confirm traf- fic incident during morning and evening rush hours. 3. If you are using personal mobile devices (such as cell phones or smart phones) as a data source, please describe the penetration rate of this data source. This information is not available to TCI due to data from mobile devices are indirectly obtained through business partners that TrafficCast doesn’t have information about the penetration rate. 4. Describe your company’s plans for future vehicle or per- sonal technology and data product development. TCI has the opportunity to take a close look at the mobile data collected by telecommunication carriers and is in

114 purpose in every 15-minute interval for each day in the week and by season. Its basic attributes include season, day of week, time of day, TMC segment ID, and speed. This data is updated annually and being used by some public agencies and navigation device manufactures. • Incident: gathered from public agencies and by TCI NOC (National Operation Center in both Philadel- phia, PA, and Madison, WI), the content of traffic incident data include accident, planned roadwork, and emergency events such as flooding, hurricane evacuation, etc. This information has been provided to TV stations and major telematics service providers such as Google and OnStar. • BlueTOAD data: By detecting Bluetooth signal, BlueTOAD detects the MAC address of the device. With known BlueTOAD location and the time stamp a device is detected, travel time and travel speed between two BlueTOAD units are calculated. BlueTOAD is also used to assess the percentile of traffic distribution for a small network. b) Raw data frequency and accuracy: Dynaflow-real-time and Dynaflow-predictive: The raw data used by Dynaflow-real-time, Dynaflow-predictive come from many sources and each with different update frequency and quality. • Speed data from DOTs: TrafficCast has access to 45 markets of speed data collected by city or state DOTs with update frequency generally between 3 to 5 minutes. However, the technology used to collect such data imposes a challenge of regular maintenance of those detectors such that its accuracy is certainly questionable. In last few years, TrafficCast conducted two major quality assessments for DOT speed data collected from all available markets by comparing it to GPS probe data. The finding is somewhat discour- aging due to many DOT data never change through- out a day probably because of malfunctioning. • BlueTOAD data: Sold to public agencies for corri- dor and regional traffic data collection, BlueTOAD data has been integrated into Dynaflow as part of its inputs. It is one unique data source only available to TrafficCast. BlueTOAD scans passing by Bluetooth- enabled devices every 5 seconds and aggregate the data for speed calculation every minute. Its accu- racy, which is in the range of 90% to 95%, has been verified by several DOTs and independent consulting firms through ground truth. • GPS probe data: used to create Dynaflow products, GPS probe is updated in a frequency varies from one data provider to another. Some update every un synchronized traffic control mechanism such as traf- fic signal or stop sign. BlueTOAD: It mainly covers arterial roads in urban and suburban areas. The speed accuracy is about 85% to 93% measured by the ground truth done by TCI team, TCI customers, and independent consulting firms. 7. Describe how your products are packaged to specifically serve the needs of transportation planners for travel demand forecast modeling or for congestion manage- ment programs. Dynaflow-historical is produced and updated once a year based on GPS data to provide historical trend that can be utilized for traffic demand modeling and, service level assessment, and transportation planning. It is packaged in season, day of week, and time of day in 15-minute interval for each TMC for the entire U.S. BlueTOAD data has been used to provide real-time travel time delivery through variable message sign, and informa- tion to DOT web site. For planning purpose, BlueTOAD is also used by several public agencies to produce route traffic assignment percentile for a small network, which is essen- tial for roadwork impact study and the measurement of effectiveness of detour advisory. 8. For each of your specific data products (copy and repeat this question and categories as needed), please describe the following: a) Data product name and description: Fused from a variety of sources including historical road speed trends, real-time GPS probe, speed data from pub- lic agencies, and anticipated traffic impacts such as inci- dent, construction, weather and upcoming events from both public and private sources, Dynaflow provides accurate historical, real-time and forecast road speeds for the functional class 1 to 4 roads in the United States. • Dynaflow-real-time: update every minute, Dynaflow- real-time provides speed either based on NAVTEQ link ID or TMC location code depending on client’s needs. • Dynaflow-predictive: Although there are seven days of predictive speed stored in the database, at any given time, TrafficCast only allows customer come to retrieve up to 48-hour of prediction. The reason is the quality of weather forecast, the dominant factor for the long-term traffic prediction, degrades drasti- cally beyond 48 hours. To reduce the data size, this data is made available in TMC segment. • Dynaflow-historical: TrafficCast packages histori- cal traffic trend data for traffic study or planning

115 navigation. Therefore, tremendous effort has been spent to refine traffic event location by a sophisti- cated map-matching algorithm. Data lack of suf- ficient information is not provided to customer or used by traffic impact model. d) Level of aggregation or disaggregation: Internally, TrafficCast builds Dynaflow model based on NAVTEQ link identifier. Therefore, one level of aggrega- tion in spatial domain is to aggregate multiple GPS probe points fall into the same link ID within each model itera- tion cycle (which is usually one minute, the aggregation in temporal domain inside the model) to produce one single speed output for that particular link. When the speed is produced based on TMC segment to support application needs, another layer of spatial domain is performed by merging multiple link speeds to a single TMC speed. Both Dynaflow-historical and Dynaflow- predictive are based on TMC and use 15-minute as basic time interval. e) Cost structure: The cost structure varies from one customer to another and it also depends on market segment such as public agencies, media (TV station), web portal, and mobile location-based services, etc. The table below shows how TrafficCast data products are packaged and sold to differ- ent market segments. 10 to 30 seconds, and the others update every 1 to 3 minutes. It is difficult to measure raw GPS accuracy, but there are indicators, such as GPS fix time and number of satellites observed, in some of data sources that can tell if the data is reliable so Dyna flow model can decide whether they should be used as input. • Incident data: incident data from private source usu- ally update within one minute during rush hours. While generally the public sources only report inci- dent at the beginning and the end of an event, ex- cept California Highway Patrol. It is expected that the quality of incident data is not consistent across the entire country with some include detailed event description but some only provide a short phrase such as “traffic accident” or “traffic collision” such that it is almost impossible to derive traffic impact from such limited information. Dynaflow-real-time is updated every minute, Dynaflow- predictive is updated every 15 minutes, and Dynaflow- historical is updated annually. Accuracy of Dynaflow-historical is difficult to measure, but the accuracy of Dynaflow-real-time and Dynaflow- predictive is between 65% and 90% depending on market and road functional class. c) Data cleaning process: • GPS probe data: criteria used to filter GPS probe data include its location (on the roadway or inside a building), quality index (for the data sources pos- sess such information), heading (within 5 degrees compared to road geometry calculated from the digital map), and data latency (need to be within 5 minutes). • DOT speed data: TrafficCast assigns a quality index 1 to 4, with 4 is the highest quality, to each DOT sen- sor dynamically based on offline data analysis, real- time speed trend (to make sure data changes with time),and comparison to GPS probe data. Quality index lower than 3 is not used by TrafficCast’s traffic models. • Incident data: The quality of incident data is inconsistent due to they come from different sources. Among three key attributes of an incident—time (when), location (where), and what happens, loca- tion is usually the most challenging one to deal with. For the data to be provided to customers and be used by traffic impact model, knowing the exact location of incident data is crucial (so a ramp closure won’t be mistakenly treated as a highway closure). This is particularly true for the data used to support mobile Product\Customer public agency media web portal mobile LBS Dynaflow-real-time city or state city nation wide nation wide Dynaflow-predictive state - nation wide nation wide Dynaflow-historical state - nation wide nation wide incident - city nation wide nation wide BlueTOAD city, county - - - • Dynaflow-real-time: For nationwide web portal and mobile LBS, TrafficCast charges client a lump sum annual loyal fee plus a small monthly subscription fee for each end user. If data is not use to support nationwide service, it is mainly charged based on the number of markets and the tier of market (top 20, 20–40, 40 after, etc.) they belong to. For public sector that requires statewide coverage, the price is based on total road mile. • Dynaflow-historical: It is charged based on number of road mile and data update frequency. • Dynaflow-predictive: It is charged based on the tier of market (top 20, 20-40, 40 after, etc.) and number of markets. If it is for nationwide traffic service, then

116 DOT speed: Generally, it is only available for the urban Interstate freeway system. Arterial may not be covered. Incident data: Often time data from state DOTs or 511 systems doesn’t cover traffic event occurs at local arterial. Therefore, the impact (delay and queue length) on minor roads may not always be available to Dynaflow. 10. Describe any data usage clauses of agreements that come with new vehicle/device purchases that enable or autho- rize your firm to use personal mobility data. TrafficCast does not own any product or application to collect mobility data directly from end users. 11. Describe how data source privacy and location/time of day details are protected. Privacy is a major concern to those GPS probe data provid- ers; therefore, before data made available to TrafficCast, their unique identifiers (either, vehicle ID, device ID or device series number) have been converted to a set of random numbers using hash function that is unknown to TrafficCast. the pricing structure is similar to that of Dynaflow- real-time. Incident data: It is charged by the tier of the market for media. For web portal and mobile LBS, TrafficCast charges a lump sum annual loyal fee plus a small monthly subscription fee for each end user. BlueTOAD: it is sold by number of BlueTOAD units. Data is included. Separate charges for solar panel and/or wire- less communication fee are applied if power and/or Ether- net connection is not available at the site of installation. 9. Describe any potential demographic bias that exists in your data sources. GPS probe data: the bias could be (1) it mainly comes from fleet (delivery truck and long-haul truck), (2) depending on data providers, the data could be limited to a certain region. Hence, data from multiple vendors is required to create a solid nationwide flow database, and (3) some data is collected from handheld devices, which is limited to those smart phones with GPS tracking capability enabled.

Next: Appendix D - Experiment A Models »
Applying GPS Data to Understand Travel Behavior, Volume I: Background, Methods, and Tests Get This Book
×
 Applying GPS Data to Understand Travel Behavior, Volume I: Background, Methods, and Tests
Buy Paperback | $75.00
MyNAP members save 10% online.
Login or Register to save!
Download Free PDF

TRB’s National Cooperative Highway Research Program (NCHRP) Report 775: Applying GPS Data to Understand Travel Behavior, Volume I: Background, Methods, and Tests describes the research process that was used to develop guidelines on the use of multiple sources of Global Positioning System (GPS) data to understand travel behavior and activity. The guidelines, which are included in NCHRP Report 775, Volume II are intended to provide a jump-start for processing GPS data for travel behavior purposes and provide key information elements that practitioners should consider when using GPS data.

READ FREE ONLINE

  1. ×

    Welcome to OpenBook!

    You're looking at OpenBook, NAP.edu's online reading room since 1999. Based on feedback from you, our users, we've made some improvements that make it easier than ever to read thousands of publications on our website.

    Do you want to take a quick tour of the OpenBook's features?

    No Thanks Take a Tour »
  2. ×

    Show this book's table of contents, where you can jump to any chapter by name.

    « Back Next »
  3. ×

    ...or use these buttons to go back to the previous chapter or skip to the next one.

    « Back Next »
  4. ×

    Jump up to the previous page or down to the next one. Also, you can type in a page number and press Enter to go directly to that page in the book.

    « Back Next »
  5. ×

    To search the entire text of this book, type in your search term here and press Enter.

    « Back Next »
  6. ×

    Share a link to this book page on your preferred social network or via email.

    « Back Next »
  7. ×

    View our suggested citation for this chapter.

    « Back Next »
  8. ×

    Ready to take your reading offline? Click here to buy this book in print or download it as a free PDF, if available.

    « Back Next »
Stay Connected!