A couple of weeks ago I did one of the most intense training days of my career. Luckily the trainers gave us ample breaks, wonderful healthy food and plenty of encouragement so at the time it didn’t seem that intense. It’s now I’m back at work and reflecting on those nine hours trying to work out how to apply my new knowledge that I realise how much of that learning I have to rationalise and consolidate with what I already know and do.
Decoded’s Code in a Day course came highly recommended by a colleague at Socitm but I could never justify attending as we have in-house developers. Imagine my glee then when they started a Data in a Day course and my gratitude to my boss for being allowed to go.
The morning started with a brief introduction to data and its history which included coxcomb charts devised by Florence Nightingale in 1857 to show the causes of death, by month, during the Crimean War. Although these charts are misleading – the data maps to the radius of each wedge, not the area – whether intentional or not, they helped her make her case that soldiers were dying of disease and not of the wounds they received during fighting. Her diagrams make us assume her conclusions without analysing the data behind them.
From there we had a look at open data sources including UK Government open data and data from Transport for London. This part struck a chord and will with most public sector organisations. TfL and the Government are too busy creating the data to also think about nifty ways to use it so they open it all up for developers to use for the apps some of us use every day. One of my reasons for wanting to do the course was to understand how open data could help reduce the number of FoI requests we deal with – maybe if the data was there in a format developers and journalists could use, people would stop phoning.
Our first practical session had us creating Google charts using datasets from the United Nations which got me thinking about the data local authorities sit on. This was my second reason for doing the course. Stats and figures swim in front of my eyes. Words are my game but complex data sets don’t work with words alone – pictures can be better (back to Florence and her dying soldiers). I’ve been doing a lot of work with Mosaic data lately and nothing tells a senior manager the story about their service users like a pen picture, complete with an idea of the family, the house they live in and what they like to do in their spare time. Sometimes an infographic can be more persuasive, not only with managers but with the public too. I’m keen for us to use infographics for things like the council budget, not only on our own website but also to go out to the local papers for them to use.
I’ve always thought that the Guardian is particularly good at infographics and I remember seeing their Government spending bubble graph and being impressed. The guys at Decoded showed us it again and then they showed us something even better. The Daily Bread is an interactive infographic which lets you see exactly what each penny of the taxes you pay are being spent on. But what’s nifty about this is that the data updates in the background so it’s always current at the front end. Imagine if you could put to bed once and for all exactly what people’s council tax pays for.
We also looked at some cutting edge visualisations some of which are simply beautiful to look at, mouse over or take inspiration from but next to useless as they stand (my opinion, not Decoded’s):
Movie Galaxies is a fisheye visualisation which shows the relationships between characters in films
US gun deaths in 2013 is an emotive visualisation with real-time data driving it
Out of sight out of mind is a visualisation of all the US drone strikes on Pakistan driven by live data maintained by the Bureau of Investigative Journalism
Tweetping is mesmerising and shows the world’s Twitter activity in real-time
Our afternoon practical session saw us using the Twitter API and Python to create charts/visualisations comparing four different Twitter accounts in real-time. I chose Westminster, North Lanarkshire, Falkirk and Fife. Had I realised we’d be looking at sentiment I’d have used @mentions but I ended up looking at the tweets the four councils were putting out rather than what people were saying about them so the sentiment was always going to be positive. I’ll have to go back into the code and change that to see what the charts come out like.
That took up most of the afternoon and, given my limited coding skills, was the most challenging bit – but I did it.
After that we looked at the fabulous D3 website. There you’ll find the code that makes a lot of data visualisation work, all open source and free.
And after all that my ears started bleeding and my head exploded. Only kidding – after that everyone on the course kicked back, had a chat then headed our separate ways.
I probably don’t have the confidence to tackle a big data visualisation myself but I now know that these fancy, interactive infographics aren’t done by smoke and mirrors and powered by hamsters on treadmills. They’re only as good as the data you start with but from there on in it’s a combination of art and science – a bit of coding know how and a designer’s eye. Luckily we have all three elements in spades at South Lanarkshire Council. I just have to get the finger out and get people to hand over their data.
I firmly believe that knowing your way around data and the stories it can tell is an essential tool in today’s communications toolbox so I’d thoroughly recommend this course and the many data journalism MOOCs that are springing up.
For those of you in the Scottish public sector I’m talking with Decoded about getting them up to Scotland for some tailored pop-up sessions. If you’re interested drop me an email at email@example.com and tell me what you’d like to be able to do with your data, the challenges you face and what you’d like the Decoded team to cover.
We all have quality data – we shouldn’t be afraid to use it.
Tune of the week