Better idea - save the data to a server, make an API (free), and let engineershelp visualize where the violence is occuring / predict where it will occur based on date/time
When you put hundreds of thousands of man-hours into digitizing an archive, there's essentially a cost/value-add. If there wasn't a value-add, why would you want this hypothetical database anyway? The crime information is public, why don't you just go compile it yourself!
Yes doctor, you just hit the "save to SQL database" right here on your plastic clipboard...