By Lydia Namubiru
Journalists in Uganda, as elsewhere, have always worked with numbers in the news. Typically, though, these tend to be numbers handed down by an agency or researcher who deals with data more regularly than them. Usually, journalists take these numbers and use them as garnish for stories, scattering them here or there in the text or bulletin. That was enough for both the media and their audiences for a long time.
Then, the internet happened.
The internet cut the value of regular text and byte stories. “Social media does that better,” critics tell said. However, the internet also gives journalists the resources that can cost-effectively take story-telling beyond those text and byte pieces.
As a journalist who is still invested in the old school depth reporting, the most irresistible of things on the internet for me are; the massive amount publicly available data and the awesome tools to help us find and tell stories with this data.
Over the last 18 months here at the African Centre for Media Excellence (ACME), we have taken 43 journalists through the first steps towards making use of these data resources. At start learners are understandably daunted. After all, the gold standard examples of data journalism usually cited are fancy but seemingly complex undertakings like the Guardian Datablog. What we learn in a matter of days is that there are actually much lower hanging data journalism fruits. Here, I will condense some quick lessons from the training into a cheat sheet for the journalist who wants to start doing some data journalism tomorrow.
- Find Data by Digging Deeper into Google Using Search Delimiters
You’ll be amazed by the variety of ready to use data you’ll find online if you are searching Google in a more sophisticated way for example by specifying the site or file type you are looking for. To find an excel spreadsheet, start your search use a search life filetype:xls <space> search term.
Most Ugandan data is however held in pdf reports published by ministries, NGOs and other agencies. So, be sure to also try the delimiter; filetype: pdf. To search within a specific site but using Google’s powerful search engine start your search by specifying the site as in this example: site://ubos.org kilometers tarmac Uganda.
- Know the local data eco-system
Who owns troves of useful data? Government agencies (think UBOS, ministries, accountability agencies like the PPDA, IGG, Auditor General, etc) as well as large non-profits and think tanks (think World Bank, UNICEF, ACODE, NGO Forum). Bookmark their data portals where they have them or build source relationships with their data people. Who is already publishing useful data? Lots of portals are springing up. Check out data.ug, budget.go.ug, ubos.org. The list is long. If a source organisation was already in the habit of, or by law compelled to, produce data heavy reports, chances are they are publishing data online or at least it in softcopy read to email it when you ask nicely or through an access to information request – whatever works. ACME itself is publishing data on our online resource centre. Get to know where to find the data you need and it won’t be one single place or individual.
- Get comfortable with spreadsheets
It must be emphasized that one need not become a data scientist before they can do awesome data journalism. There is a plethora of click-through free tools online that do the data science & give you ready to use visualizations and analyses. However, even to use these, one needs to first get fairly familiar with spreadsheets. Key first skills to aim at learning should be;
- Importing data into a spreadsheet – Almost all Ugandan data is in PDF reports. For a data journalist, that is not usable until you unlock it and get it into Microsoft Excel or any other spreadsheets program.
- Cleaning data – An easy-to-work-with spreadsheet should all data in a particular column relating to a single attribute of the data (e.g., salary amount, budget amount, name of the entities) and of the same format, i.e. all numbers, all text, all percentages. Each row should relate to a single entity. For example; one PLE candidate, one school, one district. It should also have no merged cells. Clean datasets are the only ones that can be meaningfully analysed and/or visualized even and perhaps especially if one is using an automated tool to do the analysis/visualization.
- Using simple formulas to generate simple new bits of data like totals, sub-totals and average and pivot tables to summarise large datasets. Hopefully, data will come to you as raw as possible. That way, you have more detail and therefore more options for which slice of the story to tell first, second and so on. However, you won’t want the reader to be swamped with such detail as the number of reams of paper bought by Soroti district council in Q3 of 2014. You need to summarise, group, sort data to more digestible summaries. That is why, you will use your spreadsheets. If you stick with them, you’ll in a few clicks summarise, group, sort etc and produce new audience ready data. This is not the same thing as getting a summary table in a report. The difference here is that you the journalist (not a researcher with a different agenda) got to choose what summary to produce.
- Play in the amazing world of data visualization tools
“What is your favourite data visualization tool?” someone recently asked me. It’s an unfair question. How is one supposed to choose favourites given all there is out there? Just Google “best data visualization tools” and you’ll be amazed and potentially overwhelmed. Certainly, there are some that are easier than others to start yourself on. Check out easel.ly, infogr.am, visual.ly and picktochart.com for a start in infographics. That is; poster-like options for presenting data. For TV journalists, moovly.com creates animated graphics. Cartodb.com is probably the easiest and I dare say most useful way to do maps. When you get more comfortable with presenting data visually, try out the fancier looking charts at app.raw.densitydesign.org. Timeline tools, typically give you an interactive slider, as a great option for telling stories that developed over time. Think of a story like ‘how riot events unfolded over the cause of the day’, or ‘how a landmark court case evolved over several years’. See timeline.knightlab.com for an easy start to timelines. Don’t forget that spreadsheets programs have functionalities for charts and graphs within them. In fact, these are a great place to start when trying to figure out which kind of charts make most sense for the data story you want to tell.
- Tell your manager to get more flexible with your digital publishing platforms
In speaking to editorial web managers, we found that many have unnecessarily limited use rights to the news sites they manage. For instance, after working with a web manager on an interactive timeline for the hottest evolving story that day, I found that he did not have the right to do anything with presentation on the site except paste in text and upload still pictures. We needed to embed a few lines of code into the source code of that story to display the timeline so we gave it up and he instead asked the in-house designer to draw the regular plain vanilla static graphic. Data journalism can be done on any platform but it definitely is way more fruitful on the web and impressive when done for the web. If you can get your management to allow you more flexibility in designing how content, can appear on your site, that’s more utility for your readers and therefore more traffic hence money for your management.
- Join the local communities of internet enthusiasts
How come there are so many data journalism aids online? That’s because there is a very enthusiastic global community outside of journalism that wants to see more data liberated and communicated. But what is even better than online aids, working with a real human being, who can help you not just with the fish but also show you how to fish. That is why you should join communities that bring together techies, journalists, people who hold data and people who do interesting things with data. Tap into this good nerd energy. They come in various cliques. There is the open data community which meets monthly — you can join it through data.ug. There is a local Mobile Monday chapter that meets every first Monday of the month. The tech hubs like Outbox on Lumumba Avenue as well as Hive Colab and The Hub both in Kamwokya are basically walk in places where you can meet enthusiasts who may collaborate with you. They also organize lots of social and co-working events. Bookmark their sites. Sign-up for their events. Walk-in from time to time.
Have yourself an adventurous starter journey then! When at a loss, Google it , search YouTube for a tutorial or walk into ACME and speak to a data journalism mentor.