Office of Planning, Environment, & Realty (HEP)
The issues of data quality and availability were major discussion topics at the exchange. While all involved had encountered major obstacles to obtaining data for their models, many also had found innovative ways to collect new data or use existing data from other sources.
Participants were asked, what data do you use in your planning process (not just for modeling and technical analysis)? Are there data you would like that you have trouble getting or maintaining?
Most of the discussion centered on demand-side data, including land use, household surveys and traffic counts. A common theme, was expressed by one participant as "can we make do with national data or when do we need local data? "
National sources include the National Household Travel Survey, the American Community Survey and its associated Public Use Microdata Sample, the Census Transportation Planning Products, Census Longitudinal Employment Household Dynamics data, and the Quarterly Census of Employment and Wages from the Bureau of Labor Statistics.
Private sources of employment data include REMI, Dun & Bradstreet, Global Insight, and Reference USA (a public version of Info USA). The Cheyenne MPO, located near the Colorado-Wyoming border, obtained employment data from the Wyoming and Colorado departments of employment.
A few participants mentioned the highway performance monitoring system (HPMS), and pointed to the benefits of having the MPO do the counts for the HPMS.
Traffic counts are frequently obtained from State DOTs, as is crash data. Population forecasts are also sometimes obtained from the State. However, there may be conflicting forecasts, "the Chamber of Commerce projects a higher growth rate, but the State DOT doesn't believe it." (Tom Mason, Cheyenne, WY).
Locally gathered data includes land use, employment, travel surveys, and traffic counts.
In a small MPO, it is "challenging to break socio-economic data into small zones." (Rob Kenerson. Bangor, ME). In some cases, land use data is available by parcel (often, from local assessors). Tax parcels provide a useful basis for allocating available socio-economic data to small areas, and can serve as a good basis for building zones. Parcels generally form a stable basis because although they are sometimes subdivided, they are rarely re-assembled into larger parcels.
As mentioned earlier, multiple sources of employment data are available. But all employment data sources have limitations, and the work required to assemble and verify employment data may be considerable.
A number of MPOs have conducted household surveys, but the panelists felt that such surveys are expensive. As a result, household surveys are conducted infrequently. Though full regional surveys are expensive, smaller focused surveys can shed light on travel characteristics such as major trip generators or visitor and external travel markets that may be difficult to understand without insights from local data.
Traffic counts are another source of data. They may include link counts (typically gathered over a period of several days via tubes on a road), intersection turning movement counts (typically gathered manually at intersections during the peak periods) and bicycle/pedestrian counts (also gathered manually). There is interest in using origin-destination matrix estimation techniques to synthesize more accurate O-D tables based on the traffic counts. The Thurston, WA, MPO conducted a cordon survey to develop external O-D traffic flow data.
External trips are often significant for a small MPO. Possible sources for external data include the Freight Analysis Framework (FAF), or a statewide model. A challenge is that these data sources are often not at enough of a geographic level of detail to be useful. (For example, in the FAF model an entire State might be one zone.) A new national long-distance passenger survey is not imminent, and it likely will not occur at the level of detail that small MPOs need. Expansion of statewide model results to provide external origin-destination (O-D) data has been applied, and some small MPOs have conducted their own surveys to develop O-D data, as noted in the next section.
Many small MPOs have found creative ways to overcome budget constraints, political barriers, and practical difficulties and collect relevant, quality data.
With the growing popularity of GPS-enabled smartphones, many MPOs and other organizations have sought ways to collect detailed data directly from users. One especially successful example is CycleTracks, an open-source application originally developed by the San Francisco County Transportation Authority (SFCTA) to collect bicycle travel data1. CycleTracks records cyclists' travel patterns via GPS and submits the data to a centralized database. Many cyclists have welcomed the software, as it also allows them to track and review their routes and other statistics.
The Bryan-College Station (TX) MPO was able to adapt CycleTracks from the freely available source code and implement it locally, even expanding it to modes beyond cycling. With help from Texas A&M University, the effort has become a very useful and relatively inexpensive method of data collection.
Concerns with GPS studies include privacy and possible under-representation of certain groups. For example, the elderly might be over-represented in a phone survey, and under-represented in a GPS survey.
The Thurston, WA Regional Planning Council used cameras for automatic license plate recognition in conducting their O-D survey. Follow up surveys were sent to the addresses based on the license plate data. Response rate of the survey was around 15 percent. Analysis of the thousands of survey responses was conducted using Scantron technology. Bharath Paladugu remarked that "privacy is an issue and coordination is the key...you need to notify the public and local jurisdictions via news articles, website postings, [and] memos ahead of time."
1 http://www.sfcta.org/modeling-and-travel-forecasting/cycletracks-iphone-and-android (accessed 3/26/2013)