Organisations all around the world run Freedom of Information sites based on our Alaveteli platform. There are also sites running on other codebases. Either way, we’re all working to the same goals, and we believe that each site can learn from the others: for that reason, we’ve set up the Access To Information (ATI) network, a place for people who run FOI sites to meet, chat and share experiences.
Within this framework, there’s the chance for practitioners to pair up and exchange specific skills under a mentorship arrangement — and last week, we heard from GONG in Croatia who have been learning from Georgia’s ForSet.
ForSet began life as a data visualisation organisation, and these beginnings have allowed them to communicate data from their FOI site as compelling, visual stories: skills that GONG were keen to get to grips with. Sara, project co-ordinator at GONG, told us about working with Jubo from ForSet on data visualisations — and how her learnings will change the organisation’s work going forward.
Sara explained that they agreed two main goals for this project: firstly, to improve GONG’s data visualisation skills; and secondly, to use data visualisation to promote their FOI site Imamo pravo znati (IPZ) to journalists and general users. They were successful on both counts, not only did Sara learn how to use new methods and tools; but their outputs also brought approximately 50 new users to IPZ, and two additional Pro users (Pro usage is free on the site, but numbers had been stagnant of late, so this was notable).
So, how did they go about it? The mentorship comprised four stages: data collection, analysis, storytelling and visualisation, with the last being very interconnected.
1. Data collection
This stage began with both sides brainstorming potential topics for FOI requests that would be good candidates for data visualisation. An initial set of 12 topics was whittled down to five: local referendums in Croatia; special advisors (Spads) in the Croatian government; local participatory budgeting projects; local support for youth civic education; and local financing of civil society organisations.
GONG then sent 575 requests to local and county authorities, from which they received 525 responses — a pretty good response rate, and higher that expected. They didn’t hit many problems, although several authorities asked for the requester’s ID details, and there was one ministry that cited GDPR as a reason for refusing information on Spads. This case has now been referred to Croatia’s Information Commissioner.
2. Data analysis
Jubo and Sara organised the responses they received into spreadsheets: they were looking for an angle or a story among the data, and tidying it up in this way was helpful for making comparisons. Could they find a story in there that aligned with GONG’s mission or advocacy?
By organising the data in this way, the pair could easily see which data wasn’t really strong enough to take any further, and which lent itself to storytelling and visualisation. At this stage they rejected some of the angles they’d begun with, narrowing their projects down to local referendums, Spads, and lowering the voting age to 16 for the EU elections (this last project is pending; they’ll be launching a campaign in late Spring).
3. Storytelling and visualisation
Two pieces of software were used for the visualisations: Canva and Flourish. Sara was already familiar with Canva, as she uses it to create social media graphics; but Flourish was new to her, and she says she is very happy to have these new skills under her belt.
Flourish allows you to create many different types of visualisations: you upload your data and it is fairly intuitive to create maps, charts, etc. Once these were in hand, they added a call to action for each story, encouraging people to use their FOI site and especially Pro.
The visualisations
Local referendums
For the story on local referendums, GONG had requested from each local authority the number that had taken place; the topics under discussion; their outcomes; and the number of referendums that were suspended due to not being valid for whatever reason.
They received more responses than expected, and this was also the first time nationwide data had been collected on the subject.
The first angle that GONG wanted to support with their data and visualisations was ‘Croatia needs a referendum law that recognises the importance of local democracy’.
The data showed that out of 47 local referendums that had been held, just 25 met the minimum turnout for the decision to be valid. Jubo and Sara mapped these, and paired their visualisations with the argument that a lower bar for turnout would encourage better participation in local democracy – demonstrated with human figures.
A local press outlet picked the story up, using the data to make their own story: they were the area that had had the highest number of referendums, so that was their focus.
Special Advisors
The FOI requests returned the names of Special Advisors, the areas they were in charge of, and the fees they were receiving. As Sara explained, in Croatia Spads are not straightforwardly employees of the government, but they have a lot of influence, and in some cases receive huge fees.
It became clear that there are two different types of advisors, under two laws; while each type has different functions, both are called Spads. First, there are external advisors who may or may not receive compensation; and secondly there is another class of Spads who are employed internally. Neither is subject to Croatia’s legislation on conflict of interest.
A pie chart was put to service to clearly show how much compensation Spads had received. This varied widely from Spad to Spad, but the criteria dictating who received how much is still unclear: it appears to be at the discretion of the individual minister.
In collecting this data, GONG unexpectedly unearthed a scandal, as they revealed one Spad who was abusing public funds. He was fired, along with the minister concerned; this resulted in nationwide coverage for the data; albeit again with the media’s own preferred focus.
Lowering the voting age
Sara says that it was a lot of work to find data to support the argument for lowering the voting age to 16 in Croatia. They wanted to show that, while young people see voting as the most efficient political action, it is denied to a large portion of them.
Proving the absence of something is always tricky, and in this case they were uncovering that there isn’t any research to show that 16 year olds lack the cognitive abilities to vote responsibly. So they focused on other angles: in some EU countries, 16 year olds can vote, and they demonstrated that those countries are doing well in democratic processes: they score highly in the democracy index and have good voter turnout.
Like many countries, Croatia’s population is ageing, so the young are in danger of being totally ignored. GONG plan to share their findings on social media in a simplified form with graphics cards, and a call to action to show support for the campaign.
Questions and answers
Once Sara had finished her presentation, members of the audience were invited to ask questions.
Q: How did GONG and ForSet work together?
A: At the beginning, they had lots of online video calls, and later on when the data had come in, they communicated a lot via comments on the spreadsheets.
Q: It feels like each of these the topics would be applicable almost everywhere: perhaps it will spark other people’s interest to do something similar for their own country. Any advice if so?
A: The questions asked in the first two sets of FOI requests were straightforward, which led to straightforward answers. The third topic was less so; Sara and Jubo had to go through lots of reports, and often the data from one source contradicted another. Also, an uncontentious topic is likely to result in more responses: something like referendums is politically neutral, unlike spads where the authorities may have reasons not disclose information.
Q: When you put the requests together, were you already thinking about the format it would be best to receive the data in?
A: In that respect, the best question is one with a yes/no answer. The reason for excluding many of the initial topics at a later stage was that the answers varied so widely that it was hard to pull out statistics or a simple throughline: you’d be comparing apples with pears. So, for example, when asking how much of a local authority’s budget is reserved for supporting civic education, and how civic education is delivered, the responses could range from “We are in favour of civic education, but leave it to schools”, to “We provide money for civic education and produce our own textbooks”. Meanwhile, some authorities wrote two pages of waffle in the free text box.
Q: Did you narrow down the topics before or after you had submitted the FOI requests?
A: Both. There were 12 topics at the start; they decided which of them were best suited to FOI, then sent requests for five of them. One the answers had been received, they narrowed it down to three.
Q: Could one make data visualisation about the other two? It’s hard to find ways to show that there’s no information. Saying that 80% of authorities don’t reply is not a very exciting way of showing things.
A: While it might not fit in with the initial aim of the project, this sort of thing can be a great way to show how well or badly FOI is working in your country. Journalists often can’t get the information they need, so build stories around the fact that there’s no data about such an important issue.
Q: We’ve seen how much GONG has benefitted from this mentorship. What, if anything, did ForSet get from this?
A: Sara was so quick and flexible, she was great to work with. ForSet also learned from the project: for example, that it is better when requesting a large amount of data, that is sorted by the public institution, so it’s easier to work with. You can request it sorted in the way that you need for your story, which might be different from how it is in public.
Also, Canva is such a great tool for visualisations. They’ve now merged with Flourish, so the have advanced data visualisation features. You just have to make sure you choose the right format: the type of charts or graphs that will show your findings the most clearly.
Finally, ForSet didn’t know about the topics that Sara suggested, so there was plenty to learn there, plus it was great to see the ways GONG employ to publish their stories on both social media and mainstream media.