"Duke hasn't defeated Carolina at home on a rainy Thursday since 1872!" - some sportscaster, probably
"The Democrats haven't won the White House without winning Iowa in a year when someone named Obama wasn't running since 1824." - some pundit, probably
These claims are stupid. They have no predictive power. Let's mock them.
- miles traveled (RITA file in data, http://www.rita.dot.gov/bts/sites/rita.dot.gov.bts/files/publications/national_transportation_statistics/html/table_01_40.html)
- updated vegetable figures?
- add airline delays, amtrak stations served, oil prices (http://www.eia.gov/dnav/pet/pet_pri_spt_s1_a.htm)
- add categorical data sets (how does this work??) {:s => "weather on election day", :v => 'is', :o => "rainy"} {:s => "the Super Bowl winner", :v => 'is', :o => "the Patriots"}
- state based "without winning Iowa or Pennsylvania"
- MAYBE: add support for candidate based exceptions ("Without a Clinton on the ballot, ")
- boolean type (e.g. "had a trade deficit", "was warmer than usual")?? ( already in correlates.yml )
- can I make a hash that represents a whole sentence?? (then isolate linguistics stuff)
- Do I even need this Noun class? Will the Lexicon thing deal with noun number for me?
- Replace binary logic with ternary, so we can deal with both-houses-of-Congress control, which can be R, D or Split
- replace units prefix/suffix thing with a template string, to allow to specify something like "$7.55/oz"
- find or make a better oil price (Fred's is 1940s-2013, and another is 1986-pres, neither of which is great)
- are prices truncated/rounded yet?
- TODO: readd commaification
Fred is a great source for data to feed to PunditBot: https://alfred.stlouisfed.org/category?cid=32217&et=&pageID=3&t= https://github.com/mortada/fredapi apparently could filter out datasets by dates of first (before 1980) and last (2015ish) observation