There's around 20 congressional districts that appear vulnerable, given this model of the 119th Congress I developed. I used LASSO to select a set of significant and minimally colinear features to a logistic model, leveraging 2023 Census estimates from the ACS.
https://rpubs.com/Grimace/forcasting-119th-congress
https://rpubs.com/Grimace/forcasting-119th-congress
Comments
District 10, Pennsylvania (Scott Perry 50.5%)
District 1, New York (Nicholas LaLota 55.2%)
District 1, Maryland (Andy Harris 59.6%)
District 10, Michigan (John James, 51.1%)
District 24, Texas (Beth Van Duyne, 60.3%)
District 1, New York (Nicholas LaLota, 55.2%)