Prediction: DC

My model has this upcoming DC Democratic Primary at around Clinton 88 / Sanders 12.

I think Sanders might do a bit better than that (due to the search trend being over-represented in the model), but not by a lot. The demographic changes in DC due to ongoing gentrification might also affect it (so it could be off by a couple percent by not having the income/race data for 2016 - as it hasn't been released). It is the only state/almost-state with a declining percent of Blacks (between the last two censuses). Household income is high. Clinton FB likes are very high. Besides demographics, DC might have an extremely favorable hometown advantage for Clinton and be the biggest landslide since Sanders in VT.

Result is Clinton 78.9%, so

Result is Clinton 78.9%, so error of 9.1%. Benchmark and Tyler didn't bother predicting this. A Predictit user, Kenton was more accurate it with their model and prediction of 75 / 25.