The news made a big buzz on 17 June in Belgium. Big Data based website Realo.com unleashed sudden polemic after months of existence (see for instance this article in French). Their mapping of real estate properties displayed not only properties for sale, but also social indicators like revenue levels, education and unemployment levels and the percentage of foreigners.
Why did the polemic start
The polemic started after the Interfederal Centre for equal opportunities found about the website and filled a claim. That’s how the whole buzz started … months after the website went live.
What the Interfederal Centre complaints about, is that Realo displays information about the inhabitants. There are of course no nominative data and what Realo does is just to indicate levels in terms of revenues, education, unemployment, origins. This seems to be a problem for the Interfederal Centre for equal opportunities although the data is publicly available.
It just takes you a few clicks on Statbel to find out how many single foreigners there are in any given area of Brussels. Do the test …
What is wrong about the use of data made by Realo ?
There is nothing really wrong about the use Real makes of the data. Actually, even an official of the privacy commission said there seemed to be no legal issues (see here this article in Dutch). So, where is the real problem ?
The real issue is that it’s all about politics.
Everyone knows that the location and many other variables command the price of any property. Not all locations are equally attractive: there is a set of objective criteria (like access to schools, public transportation, …) and there is a set of subjective criteria (is it a safe area, do your neighbors look like you so that you can socially get integrated, …). It’s basic common sense and used by everyone when looking for an apartment or a house. It’s been like this for centuries and will continue to be like this. “Birds of a feather flock together » the proverb says.
The problem is that Realo has made visible decision criteria that were until now taboo. Technically they have done nothing wrong. Using Data Mining just reveal how and which variables play a role into determining property prices. But politically, such a message it’s much more difficult to accept.
The best evidence that those variables are taken into account (consciously or not) by individuals when buying a property, is that the simulated property prices on Realo are very accurate. We did the test with 10 properties we knew the price of … and guess what … on 5 instances the selling price was exactly the price calculated by the Realo algorithm. In the worst case, the deviation was only 2.3%Tags: marketing agency belgium, real-estate, sharing economy