The first half of the peer review was a presentation by APC on specific items to the peer review panel. During these presentations, many topics came up which initiated discussion among peer review panel members and between panel members and APC. This section documents the key points that arose during these presentations.
The AMPO model is a trip-based model that uses the TransCAD platform. The panel did not think there was any reason to consider other modeling platforms given that TransCAD is widely used across the country and is the predominant software used in Louisiana.
Given that AMPO is a small MPO and the current model is a simple trip-based model, the panel suggested that AMPO focus on improvements to the current trip-based model and to possibly convert it to a more advanced trip-based model rather than consider an activity-based model. The panel suggested that AMPO could consider developing a more disaggregate model by modeling individual people rather than aggregate households at the TAZ level.
APC is interested in evaluating transit alternatives and modeling non-motorized travel. However, less than two percent of daily trips are by transit. Lafayette Transit System, the transit agency in APC, does not have interest in using the MPO model or working with the MPO for transit planning. Lafayette Transit System does not currently collect APC counts, but they are in the process of conducting a transit survey.
Similarly, APC is interested in modeling non-motorized travel. APC does have some non-motorized counts and they have looked into using Strava data which collects non-motorized data via a Smartphone or GPS device from its members. Aggregated non-motorized origin-destination and travel time data can be purchased from Strava.
Given the low transit and non-motorized shares in the region, and the lack of comprehensive data that can be used for estimation, calibration, and validation, the panel suggested not to incorporate a mode choice modeling component into the trip-based modeling system. The panel instead suggested using off-model and GPS techniques to develop forecasts of transit and non-motorized trips. The panel also suggested for APC to look into conducting a non-motorized travel survey.
APC is interested in developing TAZs at the block level, which would increase the number of TAZs from about 1,000 to a 3,000 to 5,000 zone system. Given that run times are not an issue with this model (i.e. current run times are twelve to twenty minutes depending on convergence factor used), the panel supported the desire to decrease zone size. However, they cautioned that it may be hard to develop employment and socio-demographic forecasts at such a fine zonal level and suggested discussing the feasibility with APC’s demographer consultant and then developing the most disaggregate set of TAZs for which forecast year employment and demographics can be reliably developed. They also recommended that when TAZs boundaries are created that they use census geography so that the TAZs can be aggregated to the block group level.
The panel stated that APC should not be concerned about the model run times as they can be decreased by optimizing the TransCAD programming.
APC validated its 2010 model to traffic count data. Currently APC obtains all of its traffic count data from DOTD. However, given the extensive factoring process that is applied to the counts, APC has expressed concern that the factored counts are underrepresenting actual volume. APC recently purchased their own traffic counters and plans to collect counts on key corridors. They also recently developed an extensive set of screenlines and plan to use the DOTD and APC counts to support the screenline analysis.
The panel recommended asking DOTD to provide time of day directional raw (i.e. non-factored) counts. These counts should then be reviewed closely to ensure they are reasonable. For example, since the APC economy is driven by the energy sector (i.e., price of oil), APC should expect to see a decrease in traffic volume starting in 2014.
APC has expressed interest in purchasing “big data” of travel time, speed and OD data. APC specifically looked into purchasing INRIX data but was discouraged by the high cost that was quoted.
The panel stated that APC should validate their highway assignment to roadway speed by time of day but stated that they did not necessarily need to purchase their speed data from INRIX or another company. Instead, the panel suggested that APC communicate with their congestion management project group to assess the roadway speed data available, as well as drive their region on key corridors during different parts of the day and check speeds or use Google maps to obtain speeds between different points in the region. They also suggested APC encourage DOTD to purchase travel time data for the entire state so that the cost to APC is minimal.
The panel stated that cell-phone OD data by purpose should be used with caution. The panel was much more supportive of using locally collected Bluetooth data.
The last household travel survey conducted for the region was the 1997 Baton Rouge Personal Transportation Survey. The panel recommended reviewing how the demographic characteristics of the region has changed since 1997, by for example, comparing the 2000 and 2010 census data. The panel recommended conducting a new household travel survey if significant changes have occurred.
