Rethinking Mobile Recommendations…

…Is it time to forget ‘nearby events’ and look at neighbourhoods?

Location-based services that emphasise recommending nearby events are doomed to failure. Here is why (more on this paper). We have recently studied the relationship between preferences for social events and geography in the large metropolitan area of Greater Boston. We sampled location estimations of one million mobile phone users in the area, combined the sample with social events in the same area, and inferred the social events attended by 2,519 residents. Upon this data, we tested six simple algorithms for recommending social events and found a couple of extremely interesting things:

  • The most effective algorithm recommends events that are popular among residents of an area.
  • The least effective, instead, recommends events that are geographically close to the area.

Conclusion: When we build and deploy new technologies in our cities, we should go back to what Jane Jacobs taught us about neighbourhoods.

Companies that fail to do that are doomed to failure. For more, here is the paper, which I’ll present at ICDM.

Call to action: We would be very happy to collaborate with mobile social-networking companies. My email:


Comments are closed.