To pinpoint alcohol-related crash hot spots we used the same methodology the Texas Department of Transportation uses to run crash data analyses on the state, county and city level.
TxDOT does not look at true driving while intoxicated convictions or charges when running drinking and driving statistics. Instead they look at data reported by the investigating officer. The crashes listed were the result of the investigating officer indicating on a crash report that he or she believed a contributing factor of the crash was that the driver had been drinking or was under the influence of alcohol. Also, anytime a driver's blood alcohol content result was greater than 0.00, or the officer listed a contributing factor of “Had Been Drinking” or “Under the Influence of Alcohol.”
This only includes alcohol involvement, not drugs.
TxDOT reiterates, “It should be noted that these are not DWI crashes, but only driving under the influence of alcohol. DWI implies that a driver was legally intoxicated (BAC .08 or greater).
We took what TxDOT does to another level.
Based on latitudinal and longitudinal fields in the crash data, we mapped each crash that met the above criteria. We then merged those plot points with a zip code map file and the computer counted how many crashes were in each zip code.
We could then do what had not been done before with statewide crash data -- “heat map” the crashes by having zip codes with greater crashes be darker than those with fewer crashes. We could also sort the list to determine which zip codes had a greater number of crashes, then zoom into those areas and see where the crashes were clustered.
In cases where the investigating officer or agency does not enter latitude/longitude coordinates on the map, TxDOT attempts to enter that information using the address location provided. When latitude and longitude information is paired with zip code files, about 14 percent of the state-wide plot points are lost due to incorrect latitude/longitude information. If you don’t see a zip code shaded on the heat map, there were either no crashes reported there or geocoded information for crashes in that area was incorrect.
In theory, some officers or police departments may be more aggressive than others about entering information indicating that a driver had been drinking prior to the crash – especially if the crash does not result in a DWI arrest. That could have some effect the overall numbers.