# README for `Earthquake impacts on traffic safety using crowdsourced and police-reported crash data - Accompanying Dataset`  
`2026-09-21`  

## Links to Dataset  
Dataset Archive Link: https://doi.org/10.21949/g94d-nt25

## Summary of Dataset  

`This contains 20 data files which support the publication 'Earthquake impacts on traffic safety using crowdsourced and police-reported crash data'.`  

## Tables of Contents  
##### A. General Information  
##### B. Sharing/Access & Policies Information  
##### C. Data and Related Files Overview  
##### D. Methodological Information  
##### E. Data-Specific Information for: `Earthquake impacts on traffic safety using crowdsourced and police-reported crash data - Accompanying Dataset`  
##### F. Update Log  

## A. General Information  

**Title of Dataset:**  `Earthquake impacts on traffic safety using crowdsourced and police-reported crash data - Accompanying Dataset`  

**Description of the Dataset:** `Research in support of regional earthquake early warning systems (e.g., the U.S. Geological Survey (USGS) managed ShakeAlert® Earthquake Early Warning System) has revealed a gap in information about the impacts of earthquakes on roadway safety. This dataset includes earthquake shaking and roadway crash data from crowdsourced and police crash report sources used in this research project. This research integrated earthquake shaking data, crowdsourced navigation application data, and police crash reports to quantify and investigate crash rates after seismic events. This study includes over 400 earthquakes with magnitude (Mw) 4.0 or above within the contiguous United States since 2017. The Modified Mercalli Intensity (MMI) measure of shaking from the USGS ShakeMap product is used in conjunction with road network geometry in areas of potentially felt shaking. The MMI estimates at locations of crash reports generated by users and crash reports from California Highway Patrol (CHP) are used to quantify the frequency of crashes in earthquake-affected regions. Crash frequency per kilometer of roadway before and after earthquakes in different MMI bins are compared visually, and the influence of shaking intensity on crash counts is assessed using regression analyses including control periods from the previous year. Most crash reports occur in areas where shaking is MMI 3.5 or below, owing to the rapid decay of shaking with distance from the fault, leading to much larger areas impacted by lower shaking intensities. There are only marginal differences before/after earthquakes between the frequencies of crowdsourced or CHP crash reports in these low MMI regions. There are a few individual instances of crashes after an earthquake in regions experiencing MMI greater than 3.5 where there are few or no crashes in similar control periods. However, the number of reports in those regions is low, less than ten for some MMI bins, so it is not yet possible to draw conclusions from the results in high MMI regions.`  

**Dataset Archive Link:** https://doi.org/10.21949/g94d-nt25

**Authorship Information:**  

>  *Principal Data Creator or Data Manager Contact Information*  
>  Name: `Emma Fox` `[Author ORCID](https://orcid.org/0009-0007-0817-8363)`    
>  Address: `USDOT Volpe Center, 220 Binney Street, Cambridge, MA, 02142`  
>  Email: `[Emma.Fox@dot.gov](mailto:Emma.Fox@dot.gov)`  

>  *Data Distributor Contact Information*  
>  Name: `Emma Fox` `[Author ORCID](https://orcid.org/0009-0007-0817-8363)`    
>  Institution: `U.S. Department of Transportation, Volpe National Transportation Systems Center`  
>  Address: `USDOT Volpe Center, 220 Binney Street, Cambridge, MA, 02142`  
>  Email: `[Emma.Fox@dot.gov](mailto:Emma.Fox@dot.gov)`    

>  *Organizational Contact Information*  
>  Name: `Emma Fox` `[Author ORCID](https://orcid.org/0009-0007-0817-8363)`    
>  Institution: `U.S. Department of Transportation, Volpe National Transportation Systems Center`  
>  Address: `USDOT Volpe Center, 220 Binney Street, Cambridge, MA, 02142`  
>  Email: `[Emma.Fox@dot.gov](mailto:Emma.Fox@dot.gov)`    

**Date of data collection and update interval:** `2017-2024`  

**Geographic location of data collection:** `United States`  

**Information about funding sources that supported the collection of the data:** `Funded by Interagency Agreement between U.S. Geological Survey and U.S. Department of Transportation, Volpe National Transportation Systems Center`  

## B. Sharing/Access and Policies Information  

**Recommended citation for the data:**  

>  `Chupp, William; Flynn, Dan F.B.; Fox, Emma J.; McGuire, Jeffrey J.; McBride, Sara K.; de Groot, Robert M.; Malwitz, Andrew` (`2026`). `Earthquake impacts on traffic safety using crowdsourced and police- reported crash data`. U.S. Department of Transportation, Volpe National Transportation Systems Center. <[DOI](https://doi.org/10.21949/g94d-nt25)> 

**Licenses/restrictions placed on the data:** This document is disseminated under the sponsorship of the U.S. Department of Transportation in the interest of information exchange. The United States Government assumes no liability for the contents thereof. To protect the privacy of subject participants and conform to the restrictions of the Institutional Review Board, raw and individual-level data will not be made available.  

**Was data derived from another source?:** `Yes`  

This document was created to meet the requirements enumerated in the U.S. Department of Transportation's `Plan to Increase Public Access to the Results of Federally-Funded Scientific Research Version 1.1`(https://doi.org/10.21949/1520559) and `guidelines suggested by the DOT Public Access website`(https://doi.org/10.21949/1503647), in effect and current as of December 03, 2020.  
 
## C. Data and Related Files Overview  

File List for the `USGS_Earthquake_Data_for_DTG.zip`  

>  1. Filename: `AlumRock_CHP_withMMI.csv`
> Short Description: `California Highway Patrol crash record joined to an earthquake event with MMI/context bins`

>  2. Filename: `AlumRock_Waze_withMMI.csv`
> Short Description: `Individual Waze ACCIDENT alert linked to an earthquake and alert-time bin`

>  3. Filename: `Area_By_EQ.csv`
> Short Description: `Earthquake MMI bin with associated geographic area in square kilometers`

>  4. Filename: `CHP_reports.csv`
> Short Description: `Earthquake-linked crash-length summary for a given id/time bin/MMI combination`

>  5. Filename: `CHPCounts_ByEarthquake.txt`
> Short Description: `Per-earthquake aggregate crash rates/counts across specified temporal windows`

>  6. Filename: `earthquakes_processed.sql`
> Short Description: `Row imported via COPY for table usgs.earthquakes_processed. Each row representes one earthquake in the study period, from 2017-2024, with data downloaded from https://www.usgs.gov/programs/earthquake-hazards/earthquakes.`

>  7. Filename: `NJNYC_Waze_withMMI.csv`
> Short Description: `Waze ACCIDENT alerts in NJ/NYC region mapped to earthquakes and time bins`

>  8. Filename: `Ojai_CHP_withMMI.csv`
> Short Description: `CHP crash record near Ojai event paired with earthquake metadata`

>  9. Filename: `Ojai_Waze_withMMI.csv`
> Short Description: `Waze ACCIDENT alerts associated with Ojai earthquake context`

>  10. Filename: `Petrolia_CHP_withMMI.csv`
> Short Description: `CHP crash record tied to Petrolia earthquake identifiers and time bins`

>  11. Filename: `Petrolia_Waze_withMMI.csv`
> Short Description: `Petrolia-region Waze ACCIDENT alerts associated with earthquake metadata`

>  12. Filename: `RidgeCrest64_CHP_withMMI.csv`
> Short Description: `Ridgecrest (M6.4) CHP crash records with earthquake linkage`

>  13. Filename: `Ridgecrest64_Waze_withMMI.csv`
> Short Description: `Ridgecrest (M6.4) Waze ACCIDENT alerts per alert/time bin`

>  14. Filename: `RidgeCrest71_CHP_withMMI.csv`
> Short Description: `Ridgecrest (M7.1) CHP crash records linked to earthquake metadata`

>  15. Filename: `Ridgecrest71_Waze_withMMI.csv`
> Short Description: `Ridgecrest (M7.1) Waze ACCIDENT alerts coupled with earthquake/time-bin context`

>  16. Filename: `ritis_alerts_mmi_nc75095651.sql` 
> Short Description: `Join of roadway crash alerts from CHP and Waze for Mendocino (M7.0) quake in 2024.`

>  17. Filename: `road_miles_categorized.csv`
> Short Description: `Per-earthquake MMI bin road-length totals split by highway vs non-highway meters`

>  18. Filename: `road_miles_categorized_CA.csv`
> Short Description: `California-only subset of road-length metrics per earthquake MMI bin`

>  19. Filename: `Waze_reports.csv`
> Short Description: `Aggregated Waze crash metrics per earthquake/time-bin/MMI combination`

>  20. Filename: `WazeCrashCounts_ByEarthquake.txt`
> Short Description: `Per-earthquake crash count summaries across multiple control/after windows`

## D. Methodological Information  

**Description of methods used for collection/generation of data:** 
```
These data sources used in this study to assess relationships between earhquake shaking and roadway safety. The data are connected primarily through geographic coordinates to produce a final dataset describing roadways impacted by earthquakes, as well as crash reports with associated ground shaking intensity measures:
```

| Data Type | Location | Provider | Information |
| :--- | :--- | :--- | :--- |
| **Raster Shaking Data** | Centered on earthquake epicenter latitude and longitude | USGS ShakeMap Product [1] | Rectangular array in raster format of ground shaking values on the MMI scale [2] |
| **Roadway Links** | All road links within earthquake raster boundaries | OpenStreetMap [3] | Geographic properties of roadways within the earthquake affected regions. OpenStreetMap data are updated more frequently and provide better coverage of actual physical roads than any federally managed roadway geometry data |
| **Crowdsourced Traffic Navigation App Alerts** | Latitudes and Longitudes where alerts are reported by drivers | Waze for Cities, in partnership with U.S. Department of Transportation (USDOT) [4] | Alert types, locations, and other information for alerts reported by users of the application |
| **Police-Reported Crash Data** | Latitudes and Longitudes where crashes are reported by police officers | California Statewide Integrated Traffic Records System (SWITRS) [5] | Crash records reported by CHP statewide, including location, severity, and time of the crash, among other information |
| **California Department of Transportation (Caltrans) Crash Reports and Crowdsourced Alerts** | Latitudes and Longitudes where crashes or alerts are reported | Regional Integrated Transportation Information System (RITIS) run by the Center for Advanced Transportation Technology (CATT) laboratory at the University of Maryland [6] | Alert types, severity, alert timing, and other alert-specific information for the Caltrans Crash and California Waze databases |

```
1.	U. S. Geological Survey. Advanced National Seismic System (ANSS) Comprehensive Catalog [Internet]. U.S. Geological Survey; 2017 [cited 2025 Sep 26]. Available from: http://earthquake.usgs.gov/earthquakes/search/ doi:10.5066/F7MS3QZH
2.	Wood HO, Neumann F. Modified Mercalli intensity scale of 1931. Bull Seismol Soc Am. 1931 Dec 1;21(4):277–83. doi:10.1785/BSSA0210040277
3.	OpenStreetMap Contributors. OpenStreetMap [Internet]. [cited 2025 Sep 30]. Copyright and License. Available from: https://www.openstreetmap.org/copyright
4.	Driving Directions & Traffic Reports by Waze [Internet]. [cited 2025 Sep 30]. Waze for Cities: Real-Time Traffic Data for Smarter Urban Planning. Available from: https://www.waze.com/wazeforcities/
5.	California Highway Patrol. Statewide Integrated Traffic Records System (SWITRS) [Internet]. [cited 2025 Sep 30]. Available from: https://www.chp.ca.gov/programs-services/services-information/switrs-statewide-integrated-traffic-records-system/
6.	University of Maryland CATT Lab. Regional Integrated Transportation Information System [Internet]. [cited 2025 Sep 30]. Introduction. Available from: https://ritis.org/intro
```

**Instrument or software-specific information needed to interpret the data:** `Open source data science software recommended, such as Python.`  

## E. Data-Specific Information  

1. `AlumRock_CHP_withMMI.csv`
- Number of variables (columns): `12`
- Number of cases/rows: `112,492`
- Each row represents: `California Highway Patrol crash record joined to an earthquake event with MMI/context bins`
- Missing data codes: `999 (1), 9999 (1), 99999 (1)`

2. `AlumRock_Waze_withMMI.csv`
- Number of variables (columns): `16`
- Number of cases/rows: `137,120`
- Each row represents: `Individual Waze ACCIDENT alert linked to an earthquake and alert-time bin`
- Missing data codes: `<blank>` (103,018) where optional textual fields are empty.

3. Area_By_EQ.csv 
- Number of variables (columns): `3`
- Number of cases/rows: `33,453`
- Each row represents: `Earthquake MMI bin with associated geographic area in square kilometers`

4. `CHP_reports.csv`
- Number of variables (columns): `7`
- Number of cases/rows: `1,184,112`
- Each row represents: `Earthquake-linked crash-length summary for a given id/time bin/MMI combination`

5. `CHPCounts_ByEarthquake.txt`
- Number of variables (columns): `9`
- Number of cases/rows: `253`
- Each row represents: `Per-earthquake aggregate crash rates/counts across specified temporal windows`

6. `earthquakes_processed.sql`
- Number of variables (columns): `31`
- Number of cases/rows: `413`
- Each row represents: `Row imported via COPY for table usgs.earthquakes_processed. Each row representes one earthquake in the study period, from 2017-2024, with data downloaded from https://www.usgs.gov/programs/earthquake-hazards/earthquakes.`
- Missing data codes: `\\N` (381)

7. `NJNYC_Waze_withMMI.csv`
- Number of variables (columns): `16`
- Number of cases/rows: `1,041,649`
- Each row represents: `Waze ACCIDENT alerts in NJ/NYC region mapped to earthquakes and time bins`
- Missing data codes: `<blank>` (614,098)

8. `Ojai_CHP_withMMI.csv`
- Number of variables (columns): `12`
- Number of cases/rows: `164,673` 
- Each row represents: `CHP crash record near Ojai event paired with earthquake metadata`
- Missing data codes: `999 (1), 9999 (1), 99999 (1)`.

9. `Ojai_Waze_withMMI.csv`
- Number of variables (columns): `16`
- Number of cases/rows: `311,169`
- Each row represents: `Waze ACCIDENT alerts associated with Ojai earthquake context
- Missing data codes: `<blank>` (171,362)

10. `Petrolia_CHP_withMMI.csv`
- Number of variables (columns): `11`
- Number of cases/rows: `117,662`
- Each row represents: `CHP crash record tied to Petrolia earthquake identifiers and time bins`

11. `Petrolia_Waze_withMMI.csv`
- Number of variables (columns): `16`
- Number of cases/rows: `142,847`
- Each row represents: `Petrolia-region Waze ACCIDENT alerts associated with earthquake metadata`
- Missing data codes: `<blank>` (109,355)

12. `RidgeCrest64_CHP_withMMI.csv`
- Number of variables (columns): `11`
- Number of cases/rows: `165,261`
- Each row represents: `Ridgecrest (M6.4) CHP crash records with earthquake linkage`

13. `Ridgecrest64_Waze_withMMI.csv`
- Number of variables (columns): `16`
- Number of cases/rows: `683,011`
- Each row represents: `Ridgecrest (M6.4) Waze ACCIDENT alerts per alert/time bin`
- Missing data codes: `<blank>` (557,702)

14. `RidgeCrest71_CHP_withMMI.csv`
- Number of variables (columns): `11`
- Number of cases/rows: `270,481`
- Each row represents: `Ridgecrest (M7.1) CHP crash records linked to earthquake metadata`

15. `Ridgecrest71_Waze_withMMI.csv`
- Number of variables (columns): `16`
- Number of cases/rows: `1,125,845`
- Each row represents: `Ridgecrest (M7.1) Waze ACCIDENT alerts coupled with earthquake/time-bin context`
- Missing data codes: `<blank>` (846,272).

16. `ritis_alerts_mmi_nc75095651.sql` 
- Number of variables (columns): `18`
- Number of cases/rows: `736,729`
- Each row represents: `Join of roadway crash alerts from CHP and Waze for Mendocino (M7.0) quake in 2024.`
- Missing data codes: `\\N` (1,055,850)

17. `road_miles_categorized.csv`
- Number of variables (columns): `5`
- Number of cases/rows: `10,964`
- Each row represents: `Per-earthquake MMI bin road-length totals split by highway vs non-highway meters`

18. `road_miles_categorized_CA.csv`
- Number of variables (columns): `5`
- Number of cases/rows: `5,482`
- Each row represents: `California-only subset of road-length metrics per earthquake MMI bin`

19. `Waze_reports.csv`
- Number of variables (columns): `7`
- Number of cases/rows: `2,244,780`
- Each row represents: `Aggregated Waze crash metrics per earthquake/time-bin/MMI combination`

20. `WazeCrashCounts_ByEarthquake.txt`
- Number of variables (columns): `9`
- Number of cases/rows: `411`
- Each row represents: `Per-earthquake crash count summaries across multiple control/after windows`

## F. Update Log  

This README.txt file was originally created on `2026-10-01` by `Dan Flynn` (<0000-0002-2978-5257>), `Data Scientist`, (dan.flynn@dot.gov)[mailto:dan.flynn@dot.gov]  

`Insert Today's Date 2026-10-01 Here`: Original file created