House of Representatives - 2018 Election
by Lou Galli
2018 Election Forecasting based on 2016 Election Results
With the 2018 mid-terms coming up, I decided to use Python and Beautiful Soup to scrape the 2016 election results from Secretary of State
sites throughout the country. I sorted the 435 districts based on a margin of victory power ranking.
The problem with Mid-term Elections is that the turn-out is usually much lower than during Presidential Elections so it is more difficult to accurately
forecast anything using polling data. As we are all aware, polling data was deemed very volatile in the 2016 election anyway. Therefore, using actual vote totals
from 2016 seems a much more logical way to see what might happen.
This week, I heard some polsters claiming that House Districts like Kentucky-6 and Arizona-1 are in play. A polster can claim anything, but it seems unlikely that a District that
was won by more than 15 points in 2016 would swing the other way in 2018.
Any Districts that had a <= 10% power ranking I highlighted in yellow. It makes sense that these are the true "swing districts" that could go either way.
Based on 2016 real voter data, there are 28 Republican swing districts and 14 Democrat swing districts.
Looking forward to see what happens in a few days.