PL46, Streetlight Product Assessment

Abstract PL 46, Streetlight Product Assessment Oct 1, 2022 – June 30, 2023 Determine quality and value-added from Streetlight data to help KYTC determine the need for future data acquisition. Objectives: 1) Evaluate the use of Streetlight data to improve current AADT estimation. 2) Compare Streetlight AADT estimates with Kentucky statewide model assignments. 3) Compare Streetlight estimated speed distributions with HERE and INRIX. Task 1 – Evaluate the value of Streetlight AADT data (Drs. Chen and Zhang) We intend to add the AADT estimates from StreetLight as an additional input to the existing machine learning model that utilizes various facility, demographic, network, and area characteristics to estimate AADT for all roads in Kentucky. The updated model will be calibrated and validated using traffic counts. The value of StreetLight data can then be measured by the enhanced performance of the model, if any. Task 2 – Evaluate Streetlight truck volume data (Drs. Chen and Zhang) 1. Data accuracy: Compare truck data from StreetLight with data from KYTC’s classification count stations. 2. Data coverage: Compare StreetLight truck device trips with HERE truck probe counts. 3. Value added to existing applications: add truck data to existing truck AADT estimation model and measure improved model performance, if any. Task 3 – Evaluate Streetlight speed data (Drs. Chen and Zhang) We propose analyses at selected locations to answer the following questions: 1. Is the StreetLight speed data comparable to HERE speeds when aggregated at annual level? A “Yes” answer is necessary to consider using it to measure delay for SHIFT. 2. How does StreetLight’s speed data compare to HERE probe speed data in spatial coverage? 3. How does the sample size of StreetLight’s speed data compare to that of HERE data? Task 4 – Critical facility analysis (Drs. Chen and Zhang) This analysis focuses on exploring the potential value of StreetLight OD data to improving criticality assessment method. We propose to use the segment analysis function to understand the travel pattern of trips that use each of the selected segments. This can inform the appropriateness of 1-hour travel time radius used in criticality analysis. Task 5 – Streetlight data quality for traffic estimation (Drs. Stromberg and Wang with Steve De Witte, PM) a. Synthesize national study results – Dr. Wang b. Learn to use the Streetlight API to batch submit AADT calculations – Dr. Stromberg c. Get AADT values for the 110 permanent counters in Kentucky. – Dr. Stromberg d. Use the Streetlight API to estimate AADT at these 110 locations. – Dr. Stromberg e. Compare accuracy of streetlight estimates compared to permanent counters. – Dr. Stromberg f. Use many three-day subsets of the permanent counter data to investigate accuracy of three day counts on other roads to compute AADT. – Dr. Stromberg g. Use streetlight data for the three-day subsets to compare accuracy for estimating AADT. – Dr. Stromberg h. Use streetlight data AADT estimates based on a whole year to compare three-day estimates. – Dr. Stromberg i. Report on the hypothesis: Permanent counters provide better estimates of AADT then Streetlight estimates. – Dr. Stromberg j. Report on the hypothesis: A year of Streetlight data estimates AADT better than three-day counts using permanent counter data. – Dr. Stromberg k. Compare Streetlight and KY ATR data (to confirm Streetlight and TTI/NREL paper results) – Dr. Wang