/EDGAR

Data analysis project using SEC's EDGAR system.

Primary LanguageHTML

EDGAR

  • Can negative sentiment 10-k filings predict short term movements in share prices?
  • Create a package to solve this

For each 10-k filing, count the number of words listed as 'Negative', 'Positive', 'Uncertain', etc..

Look for any correlations with short-term price movements.

BACKLOG:

  • Unit tests - I think most unit tests a broken now, as they are left over from an earlier version of the system.

  • Un-merge quarterly reports - I think I might have merged quarterly 10-Ks into one txt per year. This leaves us with fewer 'date' data points to analyze against stock prices.

  • Reorganize data directory - Simplify data directory structure. Write script to create this structure from a fresh first-run.

  • Get yahoo data at similar resolution to 10-Ks - might make sense to return quarterly trends over a longer period. Or maybe monthly. Also, make file output names more descriptive; including 'monthly' in the file name, for example.

CLI - package the project smartly into an intuitive CLI.