I have a dataset of house prices over time and have broken it down into neighbourhoods. For each neighbourhood I would like to create a time series model that captures the local price movements. What is the best approach to do so that also captures correlations between neighbourhoods over time?
Add a comment
|
1 Answer
$\begingroup$
$\endgroup$
1
You'll want to employ some type of panel data analysis (regression) that will accommodate e.g. zip codes (neighborhoods) and time series. Models like GLM, GEE come to mind. Mixed effects models can also be appropriate.
-
1$\begingroup$ You can try multivariate ARIMA although that would probably not be realistic if you have many neighborhoods. You can only predict one thing at a time this way, not the correlations between many neighborhoods. I am not sure how you would show the correlation of many neighborhoods with panel data either. $\endgroup$ Commented Apr 12, 2021 at 22:36