Decoding Pedestrian Severity at Crosswalks

This repository contains the NLOGIT codes and datasets used for paper titled "Decoding Pedestrian Severity at Crosswalks using Hybrid Clustering and Random Parameter Models" using:

• Cluster Correspondence Analysis (CCA)
• Multinomial Logit (MNL)
• Random Parameter Logit (RPL)
• Random Parameter Logit with Heterogeneity in Means (RPLHM)

using Texas CRIS crash data (2017–2022).

---

Files

sample_data.csv      # Sample dataset
All_MNL.txt          # NLOGIT code for Multinomial Logit model
All_RPL.txt          # NLOGIT code for Random Parameter Logit model
All_RPLHM.txt        # NLOGIT code for RPL with Heterogeneity in Means

---

Software

The models were estimated using:

• NLOGIT 6

---

Study Objective

The study identifies context-specific factors influencing pedestrian crash injury severity at crosswalks and examines heterogeneity across different crash environments.

---

Data Source

Texas Crash Records Information System (CRIS) Years: 2017–2022

---

Authors

• Swastika Barua
• Michael Starewich
• Tausif Islam Chowdhury
• Amir Rafe
• Subasish Das

---

Corresponding Author

Tausif Islam Chowdhury sgp98@txstate.edu