Roadway Traversal and Weather Dataset for Virginia Highways (2021–2022)

Project Description This research project examines how adverse weather conditions affect traffic behavior across Virginia’s roadway network. It integrates large-scale telematics data with high-resolution, model-based weather data to understand impacts on vehicle speeds and safety-critical events. The telematics data, collected between August 2021 and July 2022, includes movement and event records from over 116 million trips, covering approximately 1.4 billion miles. These records were map-matched to roadway segments using OpenStreetMap and Valhalla, then enriched with weather data from MyRadar, which incorporates NOAA's HRRR model and HORIS satellite inputs. The study’s core goal is to evaluate how different weather types and intensities influence speed deviation from posted limits and the prevalence of safety events (e.g., ABS, ESC). Linear mixed-effects models were used to quantify these relationships, with roadway segments treated as random effects. Data Scope The dataset includes: ~470 million roadway traversal records from 116 million trips. ~30 million sampled records for modeling. Time series weather data mapped to each roadway segment at 15-minute intervals. Observational unit: roadway segment traversal during a weather interval. Data types: Numeric vehicle kinematics (e.g., speed, speed limit, speed difference) Categorical weather classification (e.g., wet-moderate, frozen-heavy) Temporal (local timestamps) Roadway metadata (e.g., segment ID, highway class) No raw event or video files are included; the data is structured and pre-aggregated. Data Specification Data specification include in documents below.