Vicinity-like iPhone app for ‘nearby people I’ve met on-line’?

Another from the crazy ideas dept….

Whilst preparing for the ‘How to build a network‘ workshop last week I got to wondering about conferences and groupings of geeks (and Normals, but they need to catch-up with our tech first).

Why is it that when I’m at a conference or event, I don’t know if anyone nearby is a person that I haven’t met in real-life but someone that I do know online?  Surely there’s an iPhone app for that…

Here’s what I want – I start (proposed duff name) ‘WhosHere’ and it tells me, via my location:

  • TF people are nearby that are Twitter friends (two way reciprocal relationship or at least I’m following them)
  • NF people are nearby that I’ve referred to on Twitter but I don’t follow or vice-versa
  • BL people are nearby who have a blog that I’ve commented on (see below for details)
  • EM people are nearby who I’ve referred to in email recently (see below for details)
  • Same for Facebook, LinkedIn etc…

Probably I can mark off people that I know well so it doesn’t keep showing them to me (or maybe they appear in a separate tab?) – I’m interested in finding out when people I don’t know well are nearby as this will help me to turn weak-ties into stronger-ties.

Tieing a location to Twitter friends is probably really easy (assuming they’re posting location info). Presumably searching for people tweeting via a location is also easy (since iPhone apps already do it).

For the blog (BL) report the iPhone app would need to talk to a service that can check the Twitter profiles of nearby people, reference their blogs (or use a social graph explorer) and determine if I have left them a comment (since I’d use my domain when commenting) or linked to their blog.  I’d love to see this in an app!

For the email (EM) report the app would need to read my email (can it do that?) and look for names or URLs that are mentioned.  From these it can do a similar lookup via nearby Twitter people as for the blog report above.  Knowing that a company or individual is nearby that I’ve referred to in an email with a friend could be really interesting.

Am I barking up a crazy tree or does this idea make some sense?


Ian applies Data Science as an AI/Data Scientist for companies in ModelInsight and in his Mor Consulting, sign-up for Data Science tutorials in London. He also founded the image and text annotation API Annotate.io, lives in London and is a consumer of fine coffees.

5 Comments

  • Sort of like a super privacy invading google latitude via dopplr then? I guess though you could use the google social graph api to find all sites that things like twitter/linkedin/whatever users who are related in some way to you and then you'd just need a way to get their locations. Your main issue is probably the fact that not everyone is going to want to let everyone know their current location or maybe at a limited resolution. At which point it's not very useful if the only people you get to know are nearby are only those who update their location info all the time...
  • Fun idea - somehow I doubt that anything like that would ever be implemented though. It would be great for me though. I tend to be a wall flower at conferences. If I don't already know someone, I'm not likely to talk to them.
  • John - it is only super-invading if you broadcast your location (which most twitter clients do by default). I'd not want it most of the time but when in a place with 200+ geeks, some of which I'd like to meet (because I've e.g. commented on their blogs or discussed them in email) then it would be nice to know that they're present, rather than hoping for the random intros in the pub later. It is just a de-randomifyer, I don't have to know their faces necessarily, just their online identities (which they're broadcasting). At conferences people already tweet with their full location (witness 'within 1 mile' activity in iPhone Twitter apps) so the location data is already there...currently we just can't use it. Vicinity and clones already exist and the location-specific tweet data is already in the wild, why can't we just hook a few more services up and see it working?
  • Ok, so lets break this down in terms of how things like Vicinity (or AroundMe) work. * Shops/Bars etc have actual locations in the real world * Yellow pages etc (e.g. directories) publish locations (e.g. postcodes) * Shops don't tend to move around so the directories are pretty accurate * You can therefore grab all of the data and you're doing a pretty easy location search... So how could this be tackled for URLs? * URLs don't have an intrinsic location in the real world (though I vaguely recall there's some spec about how to specify a pages location via a special header or similar) * You could use (for a starting point) Twitter as the stand-in for Yellow pages * Twitter has location information, so a twitter profile page has a location * You can use google social graph API to find other URLs associated with the twitter profile page (and therefore infer their location) * Doing this in real-time would be pretty hard work/require a chunk of muscle * Though for a general (xyz URL is "in" Brighton) might be good enough if only updates every few days That would let you do the same as Vicinity and find what URLs are nearby to you. The next bit would be filtering/ordering based on how connected they are to you. Again for that Twitter may be a good starting point, as you've got connection information there. You could also probably spider the web to work out extra linkage information (though again google social graph does some of that already). Given Twitter also has some API calls for finding "nearby" Tweets you could probably do some serious mashup work and actually get something sort of working without too much effort. Wouldn't need to do it as a native iPhone app, as you can request the phone's location via Javascript in the browser anyway and then have all the javascript etc cached locally so it starts up near-instantly. Could make a fun hack day project?
  • Could indeed make for a fun hack-day project and javascript-based would make it portable between all sorts of devices (assuming location can be shared or twitter logins etc enabled). Certainly finding the overlap between sets of urls linked between two twitter accounts (where twitter tells you who you're near to) should be uber-simple?