Freight Data Sharing Guidebook (2013)

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.

48 Guideline 4 – If unrestricted data is not enough, be aware that privacy concerns must be addressed. Guideline 21 – Explain clearly to stakeholders that sharing of data will support improved freight infrastructure decisions that will benefit those stakeholders. Guideline 18 – Define and articulate the benefits, goals and purpose of data sharing to stakeholders. Guideline 19 – Include a stipulation that data is a one-time use and cannot be used for any other purposes such as regulation. To illustrate how the freight data sharing guidelines have been applied successfully, the fol- lowing examples are descriptions of two of the key case study projects reviewed in this research. The bold text shows which guideline was used. The first example used 13 of the guidelines and the second used 12. 4.1 Washington State Freight Performance Measure Project The Washington State GPS Freight Performance Measures (FPM) project used data from commercial fleet management GPS devices in trucks to develop a statewide freight performance measure program. The project involves ongoing GPS-based probe truck movement data col- lected for the Puget Sound area since 2008 and for all of Washington State since 2010. The GPS data is used by the Washington State Department of Transportation (WSDOT) to support an ongoing statewide FPM program that locates and quantifies roadway bottlenecks for trucks and provides guidance to the WSDOT capital program office as they make roadway infrastructure investment decisions. Ultimately these roadway investments improve freight mobility and this is why the private sector supports this program. The FPM project illustrates a number of different approaches to supporting freight data sharing. The original effort was funded to acquire truck data to quantify truck performance. The project team originally attempted to find readily available truck unrestricted data (Guideline 4) (such as from roadside counters and surveys) but it became apparent early in the effort that data from the private sector trucking industry was the only detailed truck travel information that was available and, if this data was to be used, privacy protection for this data would be a critical element of the project. The project eventually approached a number of private commercial fleet management GPS vendors about the use of data from their systems for this program. The vendors, who were initially reluctant to share their clients’ data with the public sector, agreed to dis- cuss data sharing after the state trucking associations stated that the data would sup- port improved freight infrastructure decisions and demonstrated their support for the program (Guideline 21) and WSDOT indicated the project’s result could help to focus resources on improved infrastructure for freight mobility (Guideline 18). In addition, the DOT assured the GPS vendors that the GPS data would be used just for the freight perfor- mance measurement and would not be used for any regulatory or enforcement purposes (Guideline 19). C h a p t e r 4 Application of the Freight Data Sharing Guidelines

application of the Freight Data Sharing Guidelines 49 The use of GPS vendors circumvented direct concerns about an individual company’s business sensitive information, but the vendors still required privacy protection in the form of a NDA (Guideline 5). GPS data feeds between the vendors and Washington State were set up only after NDAs were developed to help protect the data. The NDAs stipulated or involved both legal punitive actions as well as technical approaches such as suppression of individual truckers’ names and allowing only the release of aggregate data. The NDA process, because it involved review and negotiations by attorneys and some contract modification, was time consuming and delayed the project for more than six months (Guideline 8). The NDA agreements determined that the raw GPS would go to the University of Washing- ton who has a tradition of protecting sensitive data. The University was considered a trusted third party (Guideline 14) that would only pass aggregated analyzed data to the WSDOT and other project partners. The NDA and the resulting data sharing process were only possible after a number of technical and software approaches were agreed on to protect the identities of individual trucking companies. The vendor used software to ensure the individual IDs of the truck- ing companies were scrambled (hashed) and the NDA specified access restrictions in that aggregated database information (Guideline 11) could be released and accessible to the public and to other agencies. Ultimately the GPS data will be placed in an Internet-based data sharing and visualization package that allows manipulation and analysis of the data but this dataset also only allows access to aggregated data and any raw data will be not be accessible (Guideline 12). Another important motivator for data sharing was simply that the vendors were paid for the GPS data feeds which covered their cost of data sharing (Guideline 26). The resulting contractual relationship that was negotiated facilitated data sharing set up and supported a stable longer-term data sharing relationship. The multi-year funding was made available through the Washington State legislature and the WSDOT and has supported an ongoing effort via contracts with the GPS vendors (Guideline 25). This contract framed the data shar- ing arrangement as a business relationship and helped support the GPS vendors who needed to make an internal business case that the data sharing arrangement with the public sector was a new market and revenue stream (Guideline 24). The FPM project team was fortunate in that, early in the process, they developed a solid relationship with individuals in GPS companies (Guideline 6) that were interested in new business development. This greatly facilitated the project’s progress. 4.2 Cross-Town Improvement Project (C-TIP) C-TIP is a technology application designed to improve the efficiency of cross-town dray truck- ing movements between railroads by maximizing productive moves and minimizing unproduc- tive ones (e.g., bobtails). The several phase development and deployment test began in Kansas City in 2007 after several years of planning and coordinating with the various partners. A C-TIP Concept of Operations (ConOps) was published in March 2007 and updated in July 2009. C-TIP consists of several components, including real-time traffic monitoring/dynamic route guidance for draymen; an open architecture exchange of load data and availability informa- tion among railroads, terminal operators, and trucking companies; and wireless information exchange for truckers regarding trip assignments, traffic congestion, trip status, and location. The C-TIP deployment test applied Web Services technologies and other management techniques Guideline 5 – A non- disclosure agreement can be a good tool to support a data sharing arrangement. Guideline 8 – Begin negotiations of disclosure and use restrictions on freight data as early in the process as possible. Guideline 14 – Consider the use of trusted third parties (consultants or academics) as intermedi- aries or data analysts. Guideline 11 – Consider the use of software and database tools to protect and access freight data by removing private or competitive information. Guideline 12 – Build access restrictions into the data set as an alternative to scrubbing. Guideline 26 – Be sure to include funding to cover costs of data sharing and needed agreements to protect data. Guideline 25 – Attempt to negotiate and award public sector contracts with funding for research and data collection. Guideline 24 – Explore new market opportunities with potential data providers. Guideline 6 – A stable contracting relationship with data provider can be very helpful in successful data sharing.

50 Freight Data Sharing Guidebook to truck moves between rail companies’ terminals in Kansas City in order to reduce interchange delays, reduce the number of truck trips, and improve fleet management and terminal opera- tions, for example, by reducing bobtail or empty chassis movements. The stated public objective was to reduce congestion and improve air quality by reducing or optimizing truck movements (Guideline 22). Portions of the C-TIP project were completed during 2011, and were documented in a report by an independent evaluation contractor under a separate federal government con- tract from that of the developer/integration contractor. DOT/FHWA provided overall project funding and acted as facilitator (Guideline 25) for contractors to develop the C-TIP ConOps, Business Case, and software applications and to deploy and evaluate the pilot. The integration contractor and the separate evaluation contractor used contract funding for system devel- opment and testing. (See U.S.DOT report Cross-town Improvement Project Evaluation, Final Report, February 17, 2012.) Private company partners didn’t receive funds for their work on the project; they just provided the data as part of normal business operations. State DOTs in Missouri (Guideline 13) and Kansas were public stakeholders and provided some funding. The MPO for the Kansas City area participated (Guideline 16). Kansas City SmartPort, a not- for-profit corporation that facilitates transportation in the Kansas City area, was a participant in the promotion of the project. Open source software components for data exchange and route optimization were tested in the project and are available to apply in other inland or ocean ports. Another component tested in Kansas City focused specifically on Kansas City dray carriers. This com- ponent was tested by one local drayage carrier and involved use of a commercial software package by the carrier. The system helps locate the driver who is closest to a particular move and who is bobtailing, enabling the trucking company to assign the load or empty to that driver and thus eliminate a bobtail. Previously this effort was entirely manual, with dispatch- ers attempting to locate drivers through manual methods, often resulting in unnecessary empty moves. C-TIP supply chain partners included several rail carriers (UP, BNSF, and NS), and several local trucking companies (Mid Cities, Greer, IXT, Comtrak, and Lake Country) (Guideline 21). A local traffic data collection and dissemination group, Kansas City Scout, provided traffic flow data on the highways in the Kansas City area (Guideline 16). (How- ever, Missouri DOT turned off their data feed during the pilot test, which necessitated a traf- fic data purchase from a private local supplier of data.) Project participants had a desire to capture dray truck movement and rail inbound data from carriers to better schedule truck movements and reduce trucking congestion and delay. The test data was gathered and ana- lyzed as part of a several-month deployment test. The results and lessons learned in Kansas City have been made available to two more recent federally sponsored initiatives: a drayage optimization software development and pilot in Memphis, and the Freight Advanced Trans- portation Information System (FRATIS) project being tested in three different applications in Los Angeles, Dallas, and South Florida. Getting the railroads to share their information did present a challenge in some cases. One railroad required a NDA (Guideline 5), which took three months to complete; but another didn’t require anything before providing data. It largely depended on who the contractor talked to. Railroads have always been strict about handing out data, especially to third parties. Any time a third party was involved, a letter of authorization was required (Guideline 7), even for data coming from the same railroad. As a former railroad indus- try employee, the integration contractor is trusted by participants and already under- stood how the railroads operated, which helps assure success of data sharing agreements Guideline 22 – Add value to the data and make it available to all stakeholders. Guideline 25 – Attempt to negotiate and award public sector contracts with funding for research and data collection. Guideline 13 – Place a high priority on coordi- nation and devote the needed resources to extensive coordination with public and private stakeholders. Guideline 16 – Coordinate with local or regional agencies that may have closer relationships with data providers. Guideline 21 – Explain clearly to stakeholders that sharing of data will support improved freight infrastructure decisions that will benefit those stakeholders. Guideline 16 – Coordinate with local or regional agencies that may have closer relationships with data providers. Guideline 5 – A non- disclosure agreement can be a good tool to support a data sharing arrangement. Guideline 7 – A less formal agreement to maintain confidentiality of private sector data may be sufficient.

application of the Freight Data Sharing Guidelines 51 (Guideline 14). Key project staff used to work for the railroads, which meant they already knew who to talk to in order to get the data they needed. They also knew the right questions to ask. Data were obtained through asking the railroads repeatedly, since coordination with dray truckers is typically not a front-burner issue for them. Repeated contact proved to be the only way to gain cooperation. Railroad reluctance to share information was the main challenge, in the sense that C-TIP was not regarded as a priority for the railroads. So even if there was a potential benefit for railroads, it proved hard to obtain their cooperation. There were discussions with each railroad regarding the format of data (Guideline 12); these had to be negotiated with each railroad individually. For the railroads, the contractor found that as long as they could show the benefits, the railroads would usually agree to participate (Guideline 18). The primary selling point from the railroads’ perspective would be saving money on cross-town drayage rates through a more rational system of coordinating and dispatching moves. The railroads did recognize that the main immediate benefit would accrue to the dray companies, but that this could translate into lower rates for the railroads over the longer term. As long as participation did not cost the railroads too much, and would result in reasonable benefits to the Kansas City region, they were usually willing to share. Railroads wanted to be good corporate citizens (Guideline 20). Kansas City was chosen for C-TIP because the overall size of the terminal area in terms of volume, number of carriers, etc., was considered manageable. However, the testing found that there is not enough cross-town freight volume unless the railroads completely change their operations. There was no continued use of the test suites in Kansas City following the test, but as noted above, Memphis and FRATIS prototypes are building on the test results in Kansas City. The C-TIP design and lessons learned in Kansas City should make these new projects easier to undertake and succeed. Guideline 14 – Consider the use of trusted third parties (consultants or academics) as inter mediaries or data analysts. Guideline 12 – Build access restrictions into the data set as an alternative to scrubbing. Guideline 18 – Define and articulate the benefits, goals, and purpose of data sharing to stakeholders. Guideline 20 – Publicize the cooperation amongst project partners and seek to give the project visibility to stakeholders and the public.