Recently I completed the IBM Data Science Professional Certificate on Coursera. As part of the work we were invited to use geospatial data to analyse different city neighbourhoods. As my wife and I have been thinking about buying property here in Barcelona where we live, I decided to go a bit further to see what sort of insights statisitical and machine learning techniques could provide. I ended up doing quite an extensive amount of work (many times more than the capstone required) and I don’t want to overwhelm any of you, I’m going to divide the work into seperate posts to keep things more managable. I’m also adding the full Github repo of the work at the end and if there’s interest, I can write a post with some more details as to the methodology.
Getting to know Barcelona
One of the things that has always stuck me about Barcelona is the distinct differences between areas. From the narrow cobbled streets of the Gothic district, where you can still see the scratches that the wheels of the horse drawn carts made against the walls hundreds of years ago. and where the shop front properties are small and constrained in shape and size.
All the way to the more modern Eixample district, with it’s much wider streets, regular shaped blocks, and larger more spacious shop fronts.
Is the local pub affecting property prices?
Next, in looking at the types of venue in each area, and looking at the average property prices, I wanted to see if there was a correlatioon between certain having certain venue types close to you and the property prices in that area.
Interested in knowing what types of venues in your area most correlate with a higher or lower property price? – check it out here
Identifying similar neighbourhoods
But there are other more subtle diferences – areas that cater more to foreigners, nightlife, family life and so on. I was keen to see if we can divide the city up into groups of neighbourhoods, where if you liked one of them, you’d like likeliy to like the others, just from the profile of the types of shops and venues.
Turns out we can – and the results are here: read here
Where are the best and worst bargains to be had?
If we can see how venues nearby correlate with propety price, are we able to see neighbourhoods which appear under or overpriced? Perhaps we’re able to see where there are the types of shops and services associated with more desirable districts, but property prices haven’t caught up yet.
Want to know what areas are the most over and under priced based on their shops and services nearby – COMING SOON!