The Entropy Key I talked about recently contains a temperature sensor. The temperature of the miniature device is shown in Celsius, Fahrenheit and Kelvin when I query its statistics:


I wanted to graph this temperature to determine whether the room temperature has an influence on the key. It. does: just after midday yesterday, the sun started beating into my office, and the dip at 22:00 clearly shows that the central heating switched off.

Pachube graph of my key

Instead of using RRD or similar for graphing, I opted for Pachube which I always wanted to test anyway. Pachube calls itself a real-time open data web service for the Internet of Things, and it’s basically a service to which I can submit metrics via a REST API. Submission can be formed as XML, JSON or CSV.

A Pachube account can contain many feeds, each of which holds a number of datastreams which contain datapoints. So, for example, I created a feed called “ekeytemp” with a datastream called “0” to which I submit the Celsius temperature reading every five minutes. Each datapoint has a time-stamp and a value. Feeds are protected by API keys I create and distribute to the applications that need them. I can create as many as I want, and each feed can have its own.

Pachube also supports triggers or notifications, which can send an HTTP POST request to a URL of my choice when particular conditions occur on a datastream.

What I like about Pachube is that I can also retrieve the datapoints in a variety of different formats if I want to obtain my own data (.e.g for backup purposes). A number of hardware platforms support Pachube or can be coerced into supporting it, and there is a slew of tutorials and libraries as well.

I’ve made the feed public, so you can watch what the key is doing as long as I’m still experimenting with it.