Welcome to MTLDP’s documentation!

Introduction

The Michigan Traffic Lab Data Platform (MTLDP) is a set of Python packages for building applications on connected vehicle trajectory data, traffic network models and signal timing data. It provides the data structures that describe a traffic network, a trajectory and a signal plan, the pre-processing pipeline that produces them from OpenStreetMap, vendor GPS deliveries and agency work orders, and the utilities (loading, filtering, aggregation, plotting) that research applications on mobility and safety are built on.

Features

• Parse OpenStreetMap data into a well-defined network of nodes, segments, links, movements and arterials, and derive the signalized network around the traffic signals;

• organise the work in regions with a single configuration file, and tile large regions into sub-regions that are processed independently and in parallel;

• load vehicle trajectory data from several vendor formats and match it to the network with the Fast Map Matching algorithm;

• cut the matched points into movement-level trajectories and compute their performance indices (travel time, control delay, stops, queue distance, level of service, split failures);

• parse Signal Phase and Timing (SPaT) work orders and resolve the plan in effect for any movement, date and time of day;

• aggregate, filter, plot (time-space diagrams with the signal indication) and export the results.