Zillow Zestimate | Data Skeptic

podcasts

#1

Found the discussion about how people look at predictions very very enlightening. People seemed to prefer a bullet-point number (with some implied Median Absolute Percent Error) instead of an interval. It’s also a great example of combining data sets from many difference sources into one model.

There’s also some good reference links included in this article, including jobs at Zillow. What stood out to you?