Data pruning?

[allthesignals]

Hi,

I'm using a flat data approach for a map visualization project: https://github.com/urbanarchive/qpl-name-explorer.

I'm noticing the repository is growing quite large — almost a gig in side! Yet the actual datasets run around a megabyte or two.

Is there a good example of cleaning up some of the git history? Maybe I'm doing something wrong with the way it commits data? I could probably write part of an actual that removes some of the git history. But I'm not sure that's a great idea...

Thanks for any input!