Data-driven apartment-hunting

In Helsinki, as in all capital cities, it’s quite difficult to find a reasonably priced apartment that is relatively accessible. I mean sure, if you’re willing to live further away from the city and especially if you drive then you have more choice, but neither myself nor my wife have a car and considering the state of current carbon emissions we’re not planning to get one, so we needed to find something that was cheap, good AND accessible by public transport.

Step in data-driven apartment-hunting.

We had three criteria:

1. Absolute max price 300 000.
2. Three rooms or more.
3. Max travel time to the area of Sörnäinen 30 mins (for work and hobbies), max travel time to the central railway station 30 mins (accessibility in general).

The first two criteria can be used while searching almost any real estate site, but the last one required some work, so here’s what I did:

1. I entered the first two criteria into etuovi.com (a Finnish real estate site) and got all the apartments in Helsinki, Espoo and Vantaa that fell into that bucket. I stored the JSON objects returned by the Etuovi API into a separate file. At the time of the query, there were 813 apartments that fulfilled these criteria.
2. I used the Google Geocoding API to enrich the JSON objects with coordinates.
3. I used the Google Distance Matrix API to calculate commute times from each apartment to Sörnäinen and the central railway station on a regular Monday morning.
4. I created a Jupyter notebook to display all the results on a map for easy discovery and browsing of results. All the apartments that were 30 mins or less to our desired destinations were displayed in green, all the apartments that were 30 mins or less to one destination but not the other were in blue/purple, and all the apartments that were more than 30 mins of both destinations were in red.
5. Finally I added interactivity to the data points so on click the map would display a card with basic information, one photo (to get a feel of the place) and also a clickable link to check out the original listing on etuovi.com.

So what did I discover?

1. The centre of Helsinki and the immediate surroundings look like a blast zone. There’s nothing even close that is in our price range.
2. If you go along the transport arteries that come into the city, you can find accessible places. This is especially true along the metro lines that go to eastern Helsinki and Espoo.
3. Most of Espoo is pretty inaccessible by public transport. No wonder it has a higher car per capita ratio than Vantaa and Helsinki!

But what about our apartment-hunting? Well, to our surprise we discovered that there were some apartments that fulfilled our criteria in Laajasalo, an area we did not even consider when we started looking for a flat to buy!

We booked a viewing to see a couple of apartments there. Some were ok, some were not, but in the end we found the perfect place for us: a 3 room flat, really close to the sea and the woods and wayyy below our maximum price threshold.

It took me around a day of coding during my annual holiday to do this. I would love to see a similar feature on some of the big real estate websites like etuovi or oikotie.