[Question] Trying to analyze "time series data" for a significant trend/significant changes

Question

I want to see if a change in a value over time, in my case, number of houses sold per year. I have data from 2008 to 2018. How would I analyze this? I did the "regression" in excel but I'm still confused. I've been reading about "t-test" but I'm not sure the data meets the requirement for that.

Please help, I feel so lost.

Answer 1

With regression you can model a trend but there is plenty more to do, eg arma after eliminating the trend, finding seasonalities. It’s one of the more complex fields. I suggest if you want to learn the basics to do a course somewhere.

Answer 2

It depends on how your data is structured, but you could start with a simple linear model (even though you might violate some assumptions). Download R and RStudio, set up your data in long format in Excel, save your file in csv and import it to R. Then use the built in lm function to model house\_prices \~ time. The output will tell you by how much the prices vary by time unit and if that relationship is significant. You can use a package like flexplot to easily visualize your model afterwards.

Again, this is a fairly simple analysis and there are a bunch of stuff you should/could do, but if you're in a bind, this will probably do.