It’s not every day you get to see inside an organisation as august as the Bank of England, and it’s definitely not usual to get to analyse their treasure trove of data. Or at least, not until now. In keeping with the trend for open access to data, earlier this year the Bank released a range of their datasets to the public in conjunction with the One Bank Research agenda.
One of the main goals is to open up more of their research and datasets to the public, and to thereby cultivate a wider research community outside the traditional banking sector. To kick start analytical interest from outside, they set up a data visualisation competition with a prize of £5,000, and challenged entrants to show something new or novel in the data. The brief was open-ended with one restriction: entrants must use at least some of their data.
One of the datasets was a household survey conducted yearly by the Bank. An interesting question that the Bank hasn’t explored before is whether there are underlying types of households within the survey data i.e. do households or respondents fall into distinct groups, and if so, what characteristics define them? My team-mates and I – School alumnus Hanif Samad and our friend James – have no training in economics, but we do like playing with data! We decided to form a team and enter the competition together.
The challenge was twofold: to find appropriate analytical methods to answer the question above, and to translate the findings onto a compelling visual platform. Plotting the results from a technique called multidimensional scaling, we discovered much to our surprise that respondents can be clearly divided according to how they perceived their financial security. There was even some suggestion that gender influences how worried a person is about his/her finances, irrespective of the individual’s actual financial position. We generated an interactive visualization using the shiny package in R that allows non-expert users to appreciate and explore these findings, and more, in an intuitive way.
A few weeks passed and we had all but forgotten about hexadecimal coding, when we were told we had made it to the final five and had been invited to the Bank of England HQ to present our work! Hanif flew in from Singapore the evening before and we spent the day at a hack space in Hackney finalising our presentation.
Dressed up in suits and full of curiosity, we arrived and settled into the Governor’s dining room for a three course silver service lunch. It was amazing to see the different background of those who had entered, with web designers, statisticians, economists, communication experts, and artists all present. We had a wonderful lunch discussing the current trends and the future of data visualisation and analytics, and then we were taken on a tour of the Bank of England, a building soaked in history with fascinating stories within its walls.
Then it was presentation time; each team had just five minutes to impress. The differences between how each team had approached the challenge was very clear. We heard from those who had used static graphics, some who had visualised the raw data, those who had analysed the data before visualising, and a group who had user experience at the forefront of all their design. After a brief break, the Panel announced that the well-deserving winner was Cath Sleeman. Her entry, along with those of all the other competitors, can be seen on the Bank’s website.
We got a huge amount out of taking part in the competition, and would really recommend it to anyone considering a venture into the world of competitive analytics and visualisation. Here are our top tips:
- If you are interested in digital skills, don’t be afraid to enter these competitions. Even if you don’t win or make it to the finals, it is the fastest way to get up to speed with things like coding because it forces you to prototype, handle errors and deliver a minimal working product within an intense time-frame
- Having a diverse team is immensely useful, as people with different skill sets can call out on each other’s blind spots. A statistician may see patterns in the data that others cannot, while a user experience expert may discover gaps in the visualisation that others do not see
- Have fun! Win or not, a month-long competition like this is a thrilling experience in building competence, teamwork and camaraderie.
Christopher Jarvis is a current School PhD student working on spatial analysis of cluster randomised trials. Hanif Samad is an MSc Medical Statistics School alumnus and now works for the healthcare sector in Singapore. James Arthur Cattell works for the Cabinet Office Transparency and Open Data Team.
Photo 1: Competition finalists. Christopher Jarvis (fourth from left), James Cattell (fifth from left), Hanif Samad (fifth from right)
Photo 2: Competition finalists at lunch
Photo credits: Bank of England Media Team