Data Quality | LSU Center for Analytics & Research in Transportation Safety

Projects

Data Quality

The CARTS received the Crash Data Improvement Program (CDIP) grant in the spring of 2010. Using the funds from this opportunity as a starting point for implementation, the CARTS has made significant changes to our Data Quality program in order to improve the data collected and provided. The CDIP grant focused on FIVE areas of Data Quality: Timeliness Completeness, Accuracy, Uniformity, and Accessibility. On our data reports website, Data Reports - Performance Measures, under the “Performance Measures” section, various Data Quality indicators, including timeliness of crash reporting and completeness for several fields, are available by reporting agency.

The timeliness data on the website separates agencies by electronic and paper reporting. Highlighted in yellow, are agencies with reporting deficiencies. The site captures the number and percentage of reports collected in 30 days, 60 days, and 90 days, as well as an average number of days reported.

The completeness measures for driver and vehicle data, as well as, location information are also provided by agency on the site. The Data Quality team monitors completeness, especially for crashes involving Commercial Motor Vehicles. The CARTS then contacts agencies with deficiencies in the timeliness in reporting and completeness of crash data to notify them of the problem(s) and offer assistance when possible.

For the uniformity and accuracy portions of our data quality program, the CARTS staff pulls a percentage of crash reports from each agency based on the severity of injuries involved in the crashes (fatal, incapacitating, moderate, possible/complaint, and no injury/property damage only). The data quality employees check the samples of agency crash reports for accuracy, uniformity, and completeness of seven fields identified, with the help of the DOTD: latitude and longitude, distance from the intersection, direction of the intersection, intersecting roadway, manner of collision, sequence of events, and most harmful event.

Once the reports are checked, the employee writes a synopsis noting the number of reports checked by with the percentage of incorrect fields and writes a brief explanation to explain to the agency the correct manner to fill out the fields. These synopsizes are also compared to similar reports from previous years to give the agency a glimpse of areas that have improved or those that need further improvement. A Law Enforcement Liaison (LEL) representing the Louisiana Department of Transportation and Development (LADOTD) collects the synopses and comparisons, contacts agencies, and provides data quality training based on the findings, as needed.

In addition to the completeness, accuracy, and uniformity analysis, the data quality team works on a regular basis throughout the year, to identify, evaluate, and complete or correct missing and erroneous data from the crash reports.

The CARTS continues to work to improve our data accessibility portion of the Data Quality program. In addition to the use of business intelligence to report to agencies statistics regarding their crash reports, the CARTS actively promotes the usage of data collected.

Staff members from the CARTS are active members of the Louisiana TRCC and participate in the goals of the committee for the integration of the data with other organizations.