My problem

I want to predict housing prices in a city (for an upcoming year).

My solution

Create a MLR, where average housing price is dependent and macroeconomic fundamentals (population, gdp, household income etc.) are independent.

How do I proceed with the forecasting? Do I use time series methods to forecast the independent and model the price? Or is it possible to run a regression between the modeled price and time.

I am also looking for a piece of literature related to my question.

| cite | improve this question | | | | |
  • $\begingroup$ do you know the difference between time-series forecasting and regression. In fact a better question will be, "do you even know what you want to accomplish?" $\endgroup$ – mnm Aug 6 '18 at 2:33
  • $\begingroup$ what I want to accomplish is to forecast average housing price in a city. What I thought could be an approach is to model the price usig MLR, then predict the independent variables through time-series forecasting and using them model the "future" average price. $\endgroup$ – Jan Vo Aug 6 '18 at 2:46

The keyword you're after is the Boston Housing dataset. It's a data set of house prices in boston, along with various other variables, such as local crime, house size, distance to nearby shops/motorways etc.

It's a well known example dataset, so you'll be able to find examples of all sorts of house price prediction models. Download the data set and follow some of the examples to get a handle on ways in which you can approach your problem.

| cite | improve this answer | | | | |
  • 1
    $\begingroup$ Thanks - i think it shows how to creates the hedonic model of a housing price. What I am trying to do is to create a model (MLR), where an average housing price in the city is dependent and avg. income, population, gdp, etc of the city is independent. Then I want to predict the future average price - and I though about predicting the independent variables using time series and then using these predictions in the model I created, thus modeling "future" prices. I also read, that you can run the regression of a dependent variable and set time as an independent variable. Hope it makes sense. $\endgroup$ – Jan Vo Aug 6 '18 at 3:08
  • $\begingroup$ Setting time as in independent variable is sometimes useful but it is not where you want to start as I discussed here stats.stackexchange.com/questions/225931/… reflecting on @Huber's insightful comment $\endgroup$ – IrishStat Aug 6 '18 at 12:19

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.