Build a Semi-Automated Geocoding Program for Text Documents

It would be really cool to have a program that would semi-automate the process of geocoding textual data within a document in preparation for GIS analysis.  Given the complexities of place names, such a program would require some user validation for each data item.  As far as I know, there is no free software that does this.  A really ambitious program would also be able to map the data and expose the strengths of relationships among places for any set of text documents.  My particular interest in such a tool would involve the use of oral history transcripts.  I realize this is a pie-in-the-sky proposal.

Categories: Session: Make |

About Andrew Hurley

I am a professor in the Department of History at the University of Missouri-St. Louis. My fields of interest are urban history and public history. I am a member of the team that has developed the Virtual City, a software package that reconstructs lost urban landscapes digitally in 3D.

3 Responses to Build a Semi-Automated Geocoding Program for Text Documents

  1. You might be able to use OpenCalais to parse the text and find place names, and then feed them to the Google Geocoding API.

  2. One of the more successful efforts I have seen was presented at the FOSS4GNA conference this year.

    CLAVIN is also an entity resolution engine, meaning that it can often discern which state you are talking about when it processes “Springfield”.

Leave a Reply