The Apteco Datathon:
We could repeat the process by looking into the other variables and seeing which values correspond to expensive property prices. This process also gives me some understanding of where I can improve my model categories by changing the bandings (for instance changing the Yearbuilt bandings into decades). In reality I would spend some cyprus mobile number example time exploring these relationships to improve the quality of my model and removing variables of low predictive quality so that my final model ended up with a small set of variables with good predictive power.
This model can then be used to predict whether a property is going to fall into the expensive categories before its sale based upon the other factors that are known in advance of the property being sold.
Conclusion
The size of the dataset used in this post is very small in comparison to many FastStats systems, but nevertheless it has enabled us to be able to find interesting insights about the Melbourne housing market. It would be interesting to add data about property price sales in other Australian cities to compare the trends across time and across different parts of Australia.