/peds-midblocks-intersections

Repository for data used in the paper, "Analysis of Pedestrian Crashes at Intersection and Midblock Segment Levels in Texas", under review for Transportation Research Record

MIT LicenseMIT

Repository peds-midblocks-intersections

This repository serves as the distribution source for data that is used in the paper:

Zuniga-Garcia, Natalia, Kenneth A. Perrine, and Kara M. Kockelman. Predicting Pedestrian Crashes in Texas’ Intersections and Midblock Segments. Sustainability, 2022, 14(12), 7164.

As referenced in the paper, these data consists of ~1-mile roadway segments that are derived from the 2018 version of the TxDOT Roadway Invetory, as well as street intersections that are derived from the same data source, with the assistance of OpenStreetMap.

Data Files Location

Because of file size limits that are a part of GitHub accounts, the large data files are located here: https://utexas.box.com/v/peds-midblocks-intersections.

Uniform ~1-mile Segments

Since geometry representation is a vital aspect of the uniform 1-mile segments, the 1-mile segments are provided as a ZIP file of an ArcGIS Shapefile. The non-geometric columns are also provided in CSV format. These are provided as the segs_tx_1mi table.

The fields are defined as follows:

  • road_gid: Corresponds with gid found in the TxDOT Roadway Inventory, which identifies a roadway. The linear reference along that roadway would then be identified with ref_begin. The underlying TxDOT Roadway Inventory segment that overlaps the most with this "uniform" 1-mile segment can be uniquely identified with the road_gid/frm_dfo combination. Use this to key into TxDOT Roadway Inventory to find street name, road characteristics, etc.
  • ref_begin: Linear reference (in miles) along the gid roadway that marks the beginning of the "uniform" 1-mile segment.
  • ref_end: Linear reference (in miles) that marks the end of the "uniform" 1-mile segment.
  • seg_count: Segment number counting from 1 to seg_total ordered by ref_begin.
  • seg_total: Number of "uniform" 1-mile segments that comprise the entire length of the gid roadway.
  • seg_len: Length of the specific segment (in miles), which is ref_end - ref_begin.
  • frm_dfo: Exact linear reference key (in miles) into TxDOT Roadway Inventory that represents the underlying segment that overlaps this segment the most.
  • overlap: Amount of overlap (in miles) to the underlying TxDOT Roadway Inventory segment keyed by the gid/frm_dfo combination.
  • on_system: Signifies whether the underlying roadway segment is maintained by TxDOT or if it is a county or city road.
  • center_lat, center_lon: Convenience coordinates of the center of the roadway geometry.
  • ident: A string identifier that represents the roadway (e.g. street name or highway name)

Uniform ~0.1-mile Segments

There is also another dataset with the same schema as the ~1-mile segments, but representing the TxDOT Roadway Inventory chunked into segments that are approx. 0.1 miles in length. They are provided under the name segs_tx_01mi.

Intersections

Intersections were located using the methodology outlined in the paper referred above, detailed in the peds-crash-techvol repository. Please contact Kenneth Perrine, owner of this repository, for further details. Intersections are given identifiers and geo-locations. An additional table identifies the two most frequented nearby roadway sections that cross through the intersection, labeled as major and minor approaches.

ints_tx

First, the ints_tx table is provided as a shapefile and also a CSV with latitude and longitude coordinates. This is the definitions for the fields:

  • int_id: Arbitrarily-assigned, unique identifier for each intersection
  • signal: Boolean signifying whether the corresponding OpenStreetMap intersection has a signal flag associated with it
  • junction: Boolean signifying if the intersection is actually an expressway junction (e.g. a merging point for an on-ramp). Please note that in the lastest methodology, intersection-finding around expressways is subject to a number of errors. To increase reliability for idenifying traditional urban intersections, it may be appropriate to filter out all records where junction is set.
  • signal_mid: Boolean signifying whether the intersection marks the location of a midblock signal, where the signal likely does not serve a crossing of two or more public roads. This may be set in places where a midblock crosswalk is signalized.
  • descr: A string identifier for the intersection. Normally, this would be the name of the major cross-street followed by an ampersand and the name of the minor cross-street. In many cases, this is NULL, indicating uncertainty in determining the cross-streets for the intersection. It is intended that this will be corrected in a future rendition of this data set.
  • lat/lon or center: Location estimated for the intersection

ints_tx_appch

Second, the ints_tx_appch table estimates major and minor cross-streets, provided as a CSV file that is defined with these columns:

  • int_id: Refers back to the int_id identified in ints_tx.int_id.
  • road_gid: Identifies the corresponding roadway as found in segs_tx_1mi or the TxDOT Roadway Inventory via gid
  • frm_dfo: This is used along with road_gid to identify the roadway segment in TxDOT Roadway Inventory that corresponds with the location of the intersection. See note in major below about caveats.
  • lin_ref: Linear reference (in miles) along the given roadway that is closest to the location associated with the intersection. This is NULL in the few places where disjoints in the TxDOT Roadway Inventory geometry prevented the estimation of the linear reference.
  • major: Boolean that indicates whether this is the major or minor cross-street for the intersection. Since this is a simplistic characterization, intersections that have a third cross-street are not fully represented here. Also, for divided roadways as depicted in the TxDOT Roadway Inventroy, the directional carriageway is arbitrarily chosen.