Cleaning the clutter, narrowing the focus and using a relatable narrative voice in data journalism: An interview with Laura Grant

Laura Grant is the cofounder of Media Hack Collective (MHC), an independent media company that focuses on data-driven journalism. She started MHC in 2017, with partner Alastair Otter, after working for more than 20 years in South African newsrooms, including the Mail & Guardian among others.

One of their first projects as MHC was the Passmark Initiative, which focused on producing stories using South African school data, a number of which won local and international awards.   During the Covid-19 pandemic, MHC created the popular Coronavirus in South Africa dashboard. In 2021, MHC launched The Outlier, an online data journalism publication. Work published on The Outlier includes the Unemployment Tracker, Municipality Tracker and over 100 narrative charts.

Laura says that she “kind of fell into journalism by accident. Now I’m a media entrepreneur specialising in data journalism, I suppose.” Her passion and skill at creating nuanced, authentic and credible stories make it seem like no accident. 

Here are her four top tips for those who are just starting out.

1. Learn how to find and clean data

Laura spends a lot of time “rummaging for datasets” that allow her to tell evidence-based stories that may be used to inform, drive change, or raise questions. In her 10 years of doing data journalism, she has found that datasets are hardly ever in clean, accessible formats. She often has to create datasets by extracting data from multiple PDFs, spreadsheets, and other sources. “Accessible and reusable data is still hard to find, and journalists who want to do data journalism often have to compile and hand-build datasets of interest.”

You can learn how to create structured datasets, use Tabula to scrape data from PDFs and more in the Africa Data Hub Fundamentals of Data Journalism free training videos.

2. Collaborate with people who have skills you don’t have

In Laura’s experience, data journalism is best done in a team where people with different skills collaborate to produce a story. 

“There are a lot of different skills needed to do data journalism and it’s unrealistic to expect one person to be able to do everything from collecting and analysing data to reporting a story, especially under newsroom time pressures. Then you need design skills to turn data into audience-friendly visualisations, not to mention coding skills which are becoming increasingly important,” she explains.

If you want to tell great stories, the best way to go about it is with a team of people with a range of relevant skills and subject matter knowledge. Join our ADH community to find and connect with people you want to work with. Subscribe to our newsletter to stay connected and stay informed.  

“There are a lot of different skills needed to do data journalism and it’s unrealistic to expect one person to be able to do everything"

3. Use simple visualisations that make a point

Laura’s definition of great data visualisations are those that make large, complex, and inaccessible datasets  simple to understand. "It’s unnecessary to visualise all the data that you refer to in a data story and it doesn’t enhance your story to flood your readers with all the data you found in your research,” she says. "A simple data visualisation, highlighting the most relevant data, can often be more informative than very complicated charts and tables.

The complicated work is not making a visualisation, she says “It’s finding the data you need, picking the interesting threads out of it and weaving them into a narrative. If you haven’t worked out what you want to tell and show people with your data then your visualisation won’t be very useful.”

"A simple data visualisation, highlighting the most relevant data, can often be more informative than very complicated charts and tables."

4. Aim to make data useful to people

Laura spends a lot of time sifting through all kinds of datasets to find what is really interesting and useful to share with the world. “Data stories should be about people,” she said. “Data for me is: what can it tell you about the world, and what can it tell you that's useful for people? I find that really interesting. 

“Sometimes the gaps and limitations of a dataset are where the story lies,” she said. “Some of my favourite stories have come about because good data about something important didn’t exist.”

“Data for me is: what can it tell you about the world, and what can it tell you that's useful for people?"

Laura’s favourite data stories and how she uses data in creating them

Laura is most proud of a collection of data stories MHC created to highlight critical issues facing South Africa’s education system. They remain among the best examples of data journalism in the country, and are frequently used as examples when teaching new students.

1. At the time of publication, the Department for Basic Education’s audit report on the state of school toilets in South Africa had not yet been made public. Passmark took a look at the education department’s publicly available data on pit toilets in South Africa’s schools. South Africa's Deadly School Toilets (2018), and The school toilets Limpopo forgot (2018) detail school toilet conditions. The latter was done in collaboration with non-profit organisation Section27. It includes a survey Section27 did among schools in Limpopo and interrogates inconsistencies that exist in the data from various organisations. Both publications use a combination of charts, maps and tables to visualise and communicate the extent of the sanitation problem.
2. Hidden danger: asbestos in Gauteng's schools, (2018), is a four-part, in-depth report detailing the extent of asbestos use in Gauteng schools, why it is a problem, and what challenges exist with how the Gauteng Department of Education identifies which schools need to or have replaced the asbestos structures successfully. The report uses a combination of maps (static and interactive), tables, searchable tables, a video, and charts to highlight issues and challenges. Most of the data was copied from scanned documents, or scraped from text-based pdfs.
3. Twelve years in South Africa's schools (2018), takes the reader on a journey of the South African education landscape from 2005 to 2016. The story examines how student numbers change over time and what this says about student pass rates and drop out rates. The publication uses interactive charts to demonstrate the changes in student numbers and progress over time.
4. More recently, MHC has been experimenting with ways to tell data stories in their publication The Outlier and a series of newsletters that they publish fortnightly.