meltt

In collaboration with Eric Dunford, Erin McGrath, David Backer and David Cunningham (all University of Maryland/START), Karsten Donnay developed a novel method for automated integration and disambiguation of event data called Matching Event Data by Location, Time and Type (meltt). The framework performs the data merger and disambiguation based on spatiotemporal co-occurrence and secondary event characteristics. It can account for intrinsic "fuzziness" in the coding of events, varying event taxonomies and different geo-precision codes.

Karsten Donnay and Eric Dunford released a corresponding R package that computationally efficiently implements the methodology in a user-friendly interface. The package is available on CRAN, for a detailed description of the method please refer to GitHub.