Optimise for the common case.

Tiny Data hacks and experiments by @richardjpope

May 19, 2014 at 9:49pm

Habitat - a programmable personal datastore (early progress)

Habitat is a self hosted, programmable personal-data server controlled via gherkin syntax.

So far, it stores location data reported via a proof-of-concept client:

and is programmable via a proof-of-concept browser based editor:

Background photo credit:
Johnson Cameraface

May 19, 2013 at 12:31pm

The 100


In August I’m signed up to cycle the London-Surrey 100 - a race along the olympic road race route.

It’s 104 miles long and contains 800m of assent. It will probably take about 9 hours.

My training has not been going to plan (for which read none). In fact, I don’t really have a proxy for what that sort of distance feels like.

This web app shows 3 things:

1) how many days there are to go to the day iteself
2) how many days I’ve cycled in the past week
3) how my last weeks cycling compares to the race itself.

The two boxes represent distance and assent and the bit that is filled in is how much of it I’ve done in the last 7 days.

The data comes from Strava and it is permanently displayed using an old iphone.

If you are also doing the 100 you can try the app out yourself.

January 22, 2013 at 10:17pm

Food safety ratings when you checkin on Foursquare

It uses the FSA’s open data API and attempts to match them to places on Foursquare, then uses the Pushover app for iphone and android to tell you the rating.

More work is needed to improve the matching of places between Foursquare and the FSA API (due to differences in names - like ‘Ltd’ and differences in coverage). So no guarantee you will get an alert, but it seems to work quite well.

January 6, 2013 at 3:16pm

Wikicity125 finds nearby Wikipedia articles as you travel through the landscape. You can read them as you trundle along, or save them to Instapaper for later.

New articles are automatically added to the top of the screen as you get nearby, then gradually fade out as you move away. No maps, no pins, no searching.

It is built using the brilliant Wikilocation API, jQuery Mobile and Flask.

January 4, 2013 at 3:04pm

Oyster Card backup script for ScraperWiki.com Vault →

TfL only allow access to the past eight weeks of Oyster Card travel.

It’s always pained me I can’t look back on the last year (or ten years) and see where I’ve been.

This script allows you to do a rolling backup of your travel history using a private Scraper Wiki Vault

December 30, 2012 at 3:53pm

The bicycle barometer takes data about the weather, the status of the tube lines I use to get to work, and whether my local station is open or shut.

It then reduces all that data down to a single value and displays it on a dial with a bike sign at one end and a tube sign at the other.

For example, if it is raining a bit the dial will move a bit towards the tube sign, but if the tube is suffering delays, it will move a bit back in the other direction.

Different data points get different weightings. E.g. snow is more important than a bit of drizzle; the tube station being shut trumps everything.

It is built using a Nanode and an old clock I found at a flea market. The data comes from the Met Office’s Datapoint API and Transport for London’s line status and station status API’s.

The code, instructions and design for building one are available here: https://github.com/memespring/bicycle-barometer

September 14, 2012 at 10:17am

Mapping Tube WIFI: Games Underground?

There are plenty of games and apps that make use of location to collect, check-in to or find things in the real world. They use GPS and geoIP to work out your location.

But what if you wanted to build something on The Tube (London’s underground transit system)? An app that let you collect tube stations, told you their history or even a real life version of Mornington Crescent? No GPS, no geoIP.

Since July Tube stations have started being installed with WiFi run by Virgin Media.

If there is a one-to-one mapping between tube stations and the IP address you get assigned when connected to the network then it would be possible build apps that knew your location underground.

So I’ve built a mobile web app to try and build a geoIP database of the underground. Bookmark it on your phone, and next time you are commuting you can help map the tube. In return for mapping a Tube station, you get a factoid about it.

The data everyone collects will be available here.

Note: this might not work, and yes, it is all a bit silly.

Edit (Dec 2012): It seems this isn’t possible. The IP addresses don’t seem to be fixed to stations in any reliable way.