“It’s important to graph your data. You’ll notice patterns you simply didn’t realise were there” Paul explains how visualising your IoT data can reveal hidden trends… and creates a cat-checker
Iwonder how many readers will empathise with reader Daniel Franks. He’s invested a lot of money in various bits of IoT (Internet of Things) hardware, but has been struggling to find a way to graph his data, and wondered what advice I could offer via this column.
Whether you have just a couple of temperature and humidity sensors, or a complex home-automation setup, I can’t over-emphasise how important it is to graph your data. You’ll start to notice patterns that you simply didn’t realise were there.
For example, a couple of months ago, I explained here how graphs had shown the extent to which my house heats up all by itself on sunny days. Using this, combined with weather forecasts, I’ve been able to adjust the heating schedule of my south-facing rooms such that they anticipate this solar gain. In doing so, I estimate I’ve saved around 5% or more from my gas bill – that’s not to be sniffed at, and I’d never have even realised this was possible without those graphs.
That example used Domoticz, which provides graphing as part of a bigger home-automation package, but there are many standalone charting options available too.
If you’re handy with JavaScript or Python scripts, you might like to try plot.ly – it’s free if you just want a single chart, and is relatively easy to drive. Rather than overload this column with example code, I’ll simply point you to the numerous “official” examples on the plot.ly website, along with many more that you’ll find lurking on GitHub, which is always a great resource for such stuff.
Using plot.ly you can quickly produce various types of line, bar, scatter and other charts. And there’s a version 2.0 just out that has some great new features, such as the ability to connect directly to SQL databases, and a facility to use Excel formulas in your chart definitions. It’s all powerful stuff, and can be driven by only a few lines of code.
It’s strange how some IoT devices have decent data visualisation tools built in, while others are completely bereft of graphing facilities. A good example of the former is Smappee. It’s a brilliant energy monitoring tool from Belgium, which can be adopted for both home and small-business use. I’ll be looking at it in some detail in a future column, but for now I’ll just say that its data-charting facilities are excellent, allowing you to instantly see how your energy consumption varies over time, even down to appliance level, providing some great graphical tools to help reduce your fuel costs.
On the flipside, Samsung’s SmartThings – probably the most well-known and almost certainly the most-used IoT ecosystem out there – has no built-in charting facilities at all. It collates lots of data from sensors all around your house or business, but most of the time this data falls into a black hole. It’s almost criminal.