Highway Safety Analytics and Modeling

Last Updated: July 18, 2022

Review Quotes:

“This is an excellent book on highway safety (I think a must-read book who are interested to work on traffic crash data). The book has successfully enlisted all the required theory and methods to extract inference by analyzing crash data. For many years there have been a need for a single book/material which will accumulate all information relevant Read more about review stating A single material which covers all aspects for crash data collection and analysis to traffic crash data modelling. This book fills that gap…” Sumon M. 05/18/21

“This book is the most concise way of learning about highway safety analytics and modeling techniques for novice as well as for professionals. I have found this book most useful during my highway safety course. Numerous numerical examples in the book makes it much easier to understand the text. I recommend this book as a must read to all the researchers in highway safety.” GS 11/12/21

“Very informational” Clawrence 11/12/21

“For someone in the field of traffic safety, this is a must-have book. It is a unique source that includes and covers all the aspects of traffic safety, and it is among the top safety references that I have had so far.” Reihaneh 11/23/21

Latest Information

Corrections and updates can be found further below.

Lecture notes for the Tongji University lectures can be found at the bottom of the page.

Where to buy the book

The textbook is now available on Elsevier:

https://www.elsevier.com/books/highway-safety-analytics-and-modeling/lord/978-0-12-816818-9

Textbook on Amazon:

https://www.amazon.com/Highway-Safety-Analytics-Modeling-Techniques/dp/0128168188/ (You can see inside the book at Amazon)

Appendix D – Datasets

Chapter 3

Exercise_3.1_Michigan_Rural_Two-Lane_Highways_Dataset (The file is for the state of Michigan, not Texas)

Exercise_3.3_South_Korean_Dataset

Exercise_3.4_Texas_Rural_Divided_Multilane_Highways_Dataset

Chapter 4

Exercise_4.1-4.2-4.4-4.5_Large_Trucks_Dataset

Exercise 4.3: Large Trucks Dataset (Mixed Logit) (The file was too large to be uploaded on the website)
Exercise_4.3_Large_Trucks_Mixed_Logit_Dataset_1_2
Exercise_4.3_Large_Trucks_Mixed_Logit_Dataset_2_2 (copy all columns in this file into the first file starting at column BQ)

Chapter 5

No external dataset

Chapter 6

Exercises_6.1-6.3_Texas_Undivided_Highways_Dataset

Chapter 7

Exercises_7.1-7.3_Before-After_Dataset

Chapter 8

Exercises_8.1-8.8_Texas_Sample_Intersections_Dataset

Chapter 9

No external dataset

Chapter 10

Exercise_10.1_Inductive_Loop_Detector_Dataset

Exercise_10.2 Real-time_Crash_Prediction_Dataset

Chapter 11

Exercise_11.1-11.2_Car-following_Dataset

Chapter 12

No external dataset

Corrections and Updates

July 18, 2021

Corrections to manuscript *New PDF File (This file will be updated on a regular basis as corrections are being provided)

March, 2021 Update:

This paper explains how to predict crashes (out-of-sample) using random parameters models:

Hou, Q., X. Huo b, A.P. Tarko, and J. Leng (2021) Comparative analysis of alternative random parameters count data models in highway safety. Analytic Methods in Accident Research, Vol. 30, 100158.

Note: The RP model cannot be used to predict crashes without estimating the distribution of the parameters (i.e., cannot use the mean of the parameter only). Hence, it is not suitable for applying it using the methods described in the Highway Safety Manual for example.

May, 2021 Update:

In Chapter 2, Section 2.4.2, it is important to point out that naturalistic data may suffer from self-selection bias, as the drivers who are selected to participate in such studies are likely to modify their behavior and consequently reduce the risk of crashes. The magnitude of the difference compared to the general population is currently not known. Researchers should be aware of this potential issue.