The Banality of Nearby
At SXSW there was a lot of Instagramming going on. I wanted to see what sort of pictures people were taking, but there didn't seem to be an easy way of doing it and I was too busy eating to fix it myself at the time. But now,
Localgram asks your browser where you are, then shows you recent nearby uploads. That's it. It'll look decent on the desktop, and decent on an iPhone. It might work on other things. Internally, this is a single Instagram API call, with maps and pictures on it. It's so ridiculously simple that it barely seemed worth building, which is why I built it.
Playing with this around London, I've come to two conclusions.
-
Other people's Instagram photos are really boring. Good grief. Without the filters of the phone app, it's just an undifferentiated stream of graffiti, food, and duck face.
-
It's surprising to me that so many photos are geotagged. Photos at a foursquare venue have maps on the photo page, but it seems that even photos without a first-class venue can have a location associated with them and visible through the API. The default web photo page doesn't display the location at all. I hadn't realised that photos retained their location. I wonder how many other people don't know this.. Of course, this might be because some of the locations on these photos are clearly complete rubbish..
I'm playing with Instagram2 a lot recently. I like the live nature of the photo stream. I like their API (I like OAuth2). I don't like their lack of feeds and app-only approach, but I can fix the first one of those myself.
I'm also playing with the lovely Toner maps tiles from Stamen, because I was looking for a gratuitous excuse to do something with them. I'm using modestmaps.py (thanks to some prodding from Aaron) to render out the static maps - hopefully this won't kill anything - I'm not really set up to be a maptile server here. We'll see.