Scraped nearly 34,000 sentences about prospects from NFL prospect tracker, ESPN Insider, Walterfootball, FFtools, GameHaus, Drafttrek. I've already attempted to create binary variables to indicate traits that were identified in the text.
Headers:
- Sentences = Text describing prospect
- Player = The prospect
- Year = Draft Year
- Source = Source of text
The rest of the headers are feature enginered binary variables from the sentences indicating "Good" if the prospect had the trait/was good at the trait and vice versa for "Bad".
Source prefixes:
- Walterfootball = WF_
- NFL tracker = NFLtracker_
- FFTools = FFTools_
- ESPN = ESPN_
- Gamehaus = Gamehaus_
- Drafttek = Drafttek_