I'm trying to geocode a list of crime statistics that include block-only addresses i.e.:
1900 BLOCK GRAND AVENUE
700 BLOCK REED AVENUE
3000 BLOCK MISSION BOULEVARD
and a couple thousand more for one zip code for one quarter period...
I have the city, state, zip as well.
I'm new to CARTODB and am working to add a layer of crimes to visualize along with other layers such as CENSUS tracts and another dataset of certain businesses. I have the zip, census tract and business layers added but need to add this layer of crime data.
I realize this has somewhat limited value because in a heat-map situation it would show the whole block for each individual crime. But, say I only pick out a few crimes like assaults. I would then have a low enough number where a heat map might make sense.
In any case the bigger question is I have to geo-code the block number for the address and am wondering if I can use street number ranges (i.e. 700-800) or a mid-block point (i.e. 750). The problem with both of these is the bias... i.e. the crime might have happened at the corner of 700.
I'm using the Texas A&M Geocode services (amazing service for free) https://geoservices.tamu.edu as well.