How Complete are Your City’s Streets? Evaluating the Completeness of Urban Streets Using Big Data and Computer Vision

Term Start:

June 1, 2024

Term End:

May 31, 2025

Budget:

$213,472

Keywords:

Complete Streets, Detection Methods, Mode Choice

Thrust Area(s):

Data Modeling and Analytic Tools, Equity and Understanding User Needs

University Lead:

Georgia Institute of Technology

Researcher(s):

Subhrajit Guhathakurta

The main objectives of this project are: (1) development and validation of detection methods on the presence and width of individual elements of complete streets at street level, (2) development of a numeric index and typology to rate the completeness of streets, (3) curation of a publicly accessible database of various elements of complete street data in Atlanta metro; and (4) demonstration of how the rating/typology can be used to help users easily communicate and utilize the data in planning and policy decisions. Additionally, this project will provide an interactive map dashboard of non-residential urban streets in the Atlanta metropolitan region to visualize and communicate the data with the stakeholders. Understanding the current condition of complete street networks is an imperative first step in planning and policy interventions. The presence and the conditions of the eight elements of complete street segments, as reflected in the rating system, will offer critical information about which areas in the street network should be prioritized for complete street upgrades and what specific form these upgrades would entail. The knowledge of complete street elements will also help in assessing whether investing in complete street design and construction influence people’s mode choices towards more active mobility and transit. While there are numerous elements that form the concept of complete streets such as public transit facilities, pedestrian and bicyclist accommodations, traffic calming, and streetscaping, this project focuses on the elements that determine the allocation of street space, such as sidewalks, bike lanes, and street parking, and save other non-surface objects for future studies, such as signboards, walk signals, and other fixtures.

This project plans to collect data on both the presence of complete streets elements and their cross-sectional width where applicable. This detection uses aerial and street view images together as one input. The primary data source for image data will be Google Street View and Google Maps API. By validating the detected result through the comparison with the well-established data, this project will test the potential of the methodology as a low-cost alternative to the existing data collection system. All data will be collected at the street segment level. The relationship between complete streets and travel behavior outcomes in terms of urban vitality and public health will be demonstrated by using urban vitality data (e.g., daily median spend for each POI from Safegraph) and mobility pattern (foot traffic data from ADVAN Research). All data will be collected at the street segment level.

Project scope

Scroll to Top