Produce a regression for evictions

e-n-f / housing-inventory

San Francisco housing construction history and associated data

https://experimental-geography.blogspot.com/2016/05/employment-construction-and-cost-of-san.html

134 stars 25 forks source link

Produce a regression for evictions #4

Open kevinburke opened 6 years ago

kevinburke commented 6 years ago

The blog post you wrote has a regression that can predict housing prices pretty well based on three variables: supply, wages, and unemployment.

It would be neat if we could get a similar regression that could predict the number of evictions. (It might just be the same variables but that would also be interesting to know).

e-n-f commented 6 years ago

It would be interesting. Do you have the data?

kevinburke commented 6 years ago

Oh I assumed we could use the same data that you have, but instead of predicting rent prices, predict evictions. ie. "a 1% increase in housing stock leads to a X% drop in evictions"

e-n-f commented 6 years ago

You're right. I forgot that I had already included the count of evictions per year.

Do you see a pattern that predicts it?

kevinburke commented 6 years ago

This is the point of running a regression, right? I would guess that increasing the housing supply would decrease evictions, but what do I know...

e-n-f commented 6 years ago

You can do the regression, and maybe there are fewer evictions in years with more new construction, but it seems pretty weak (R²=0.0117), so I am not eager to state a relationship between the two.