Using Twitter to Predict Crime:
Python Libraries: scikit-learn sqlite3 shapefile
Steps for replicating thesis:
- Add tweets2.db back into tweets.db
- run "python extractCrimeData.py Crimes_-_2001_to_present.csv crimes.p"
- run "python tweetsForBeats.py crimes.p tweets.db geo_aerh-rz74-1.shp beats.p"
- divide tweets.db into different db for each day, put them in directory named "DBs"
- run "python putRatingsInDb.py"
- run "python getAggression.py beats.p outcomes.p features.p"
- run "python cross_validate.py [NUM_FOLDS] features.p outcomes.p"