We need to know where the model is predicting the largest swings in AVs so that we can make sure these changes are justified. Let's map out prior_near_yoy_change_pct by neighborhood and create a table that allows us to look at the top 10 largest changes by township. I'd recommend using DT for this table since it'll allow you to use drop down column filters and select individual townships to display.
The map doesn't need to be too verbose: we're primarily looking for problem areas. In the table let's include the pin, town, nbhd, class, change for the individual PIN, the median change for the neighborhood, any recent sales the PIN had, and a few characteristics like bldg_sqft, beds, and baths.
We need to know where the model is predicting the largest swings in AVs so that we can make sure these changes are justified. Let's map out
prior_near_yoy_change_pctby neighborhood and create a table that allows us to look at the top 10 largest changes by township. I'd recommend using DT for this table since it'll allow you to use drop down column filters and select individual townships to display.The map doesn't need to be too verbose: we're primarily looking for problem areas. In the table let's include the pin, town, nbhd, class, change for the individual PIN, the median change for the neighborhood, any recent sales the PIN had, and a few characteristics like bldg_sqft, beds, and baths.