Effects of 5D built environment and non-built-environment factors on injury crash risk: An interpretable machine learning analysis
Wenfang Li, Xingchen Zhang, Sen Cao · 2026
Read Effects of 5D built environment and non-built-environment factors on injury crash risk: An interpretable machine learning analysis by Wenfang Li, Xingchen Zhang, Sen Cao (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/plos-pone-0353205.txt.
Description
Wenfang Li, Xingchen Zhang, and Sen Cao present this 2026 PLOS One study on what makes a traffic crash in Changsha, China, likely to injure someone. Using nearly ten thousand crash records from 2017 to 2019, they combined built-environment features, grouped into five dimensions such as density and transit access, with crash-context factors like lighting. Three machine learning models were trained and the best was pried open to rank its influences.
Models using both kinds of information beat models using either alone, with the combined approach discriminating best. Distance to the nearest metro station, lighting, road density, and land-use mix ranked among the strongest factors. Farther metro access generally raised injury risk, while denser, more mixed neighborhoods lowered it. The authors frame built environment and crash circumstances as partners, not rivals, in explaining injury risk.
Catalog record
| CREATOR | Wenfang Li, Xingchen Zhang, Sen Cao |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 75044 |
| LICENSE | CC_BY |
| RIGHTS | CC BY |
| SOURCE | PLOS |