Currently, Schedules starts by loading the CSV for a selected semester, where the CSVs are generated by a really ugly, OS-X-based scraper application. Someone mentioned the prospect of connecting it to a data feed from the registrar, which is better for obvious reasons. This could be either a live pull, or (probably better) a server-side generation of the CSVs.

One nuance that we'll have to handle, either way: 99.999% of the CSV is just a row of cases, but the tail end of the file also includes a few records about schedule changes for the semester - dates when classes are skipped (eg: holidays), canceled (eg: snow days), or rescheduled (eg: all classes from a particular Monday are pushed to Tuesday). That data obviously doesn't come from the registrar, so we'll have to handle it separately.

5/22 - Looking at python scraping with http://scrapy.org/ using https://realpython.com/blog/python/web-scraping-with-scrapy-and-mongodb/