详情

原文传递 Evaluating Cell Phone Data for AADT Estimation.

题名：	Evaluating Cell Phone Data for AADT Estimation.
作者：	Codjoe, J.; Ashley, G.; Saunders, W.
关键词：	Streetlytics, Location data, Transportation analysis, Privacy
摘要：	AADT estimation facilitates the process of determining and prioritizing funds for both road safety and improvement projects. Traffic counts over non-state roadways are normally the responsibility of local governments and the Metropolitan Planning Organizations of Louisiana鈥檚 major cities. Other non-state roadways in rural and small urban areas do not have such systemic traffic counts or estimation programs so these areas tend to lack timely traffic volumes across non-state and local roads. Streetlytics is being considered as a potential supplement for systemically providing AADTs on all state and non-state roadways within Louisiana. In this study, a bivariate correlation analysis and quantitative analysis (percentage difference) were used to evaluate the comparability of Streetlytics鈥?volume counts with traditional DOTD counts. Simple linear regression was then used to develop predictive models that accounted for the differences between the two datasets. The data was analyzed on three levels to compare Traditional count data to Streetlytics counts. Level 1 comprised all data, level 2 comprised Routine count data versus Permanent count data, and level 3 comprised Observed locations versus Unobserved locations. Overall, the results confirmed strong positive correlation between Streetlytics data and Traditional count data. Analysis of all the data in Level 1 produced a percentage difference value of 44.50% with Traditional count data reporting higher values than Streetlytics count data. Routine counts and Permanent counts analyzed at level 2 had percentage differences of 45.01% and 43.00% respectively. At level 3, Unobserved locations from Routine and Permanent count data had percentage differences of 53.90% and 43.00% respectively while a percentage difference of 23.60% was obtained for Observed locations (Routine count data). Furthermore, the percentage difference between the two datasets for Traditional count data under 300 vpd was 110.38%. For Traditional count data over 300 vpd however, the percentage difference between the two datasets was 37.08%. Approximately 10% of the data falls under 300 vpd and 3% falls below 50 vpd. The study recommends adoption of the On-Premises dataset of Streetlytics to supplement the efforts of DOTD鈥檚 Traffic Monitoring Unit. The study further recommends subscribing to the Monthly data which takes into account seasonal variations and provides typical volumes for a day in a month of a given year and negotiating with the vendor to set minimum AADT value to 50 vpd from the current value of 300 vpd.
报告类型：	科技报告