The art in data science
By narrating overlooked lived experiences, Dimpho Mashile looks to address the erasure of marginalised groups
Dimpho Mashile is a multidisciplinary storyteller. Sometimes her medium is visual and at other times it is numerical. Through a series of phone calls, emails and late night Whatsapp chats, the Mail & Guardian spent the last two weeks speaking to Mashile in the hope of understanding a practice that involves concepts — numbers and visuals — that are often perceived to exist on the opposing ends of the information spectrum.
Hailing from emalahleni in Mpumalanga, 25-year-old Mashile is a data science student at Umuzi and a part-time creative director. Since 2014 Umuzi has offered matriculants an alternative form of higher education. Through 12-month paid learnerships, the nonprofit in Jeppestown, Johannesburg, tackles unemployment by equipping its students with skills in the creative and technological fields.
Data science is the study of numerical statistics with the goal of finding insightful trends about people’s behaviour. Also known as data points, these statistics can be collected in a variety of ways, ranging from the information we give when we register our phones all the way to the answers we give in questionnaires.
A student in the data science programme spends nine months at the Umuzi studio where they are taught skills such as programming, data collection, visualisation and analysis. The last three months are then spent gaining work experience from one of the organisation’s partners because, “in agreeing to fund our learnerships, Umuzi’s partners get back three months worth of the skills that they invested in creating”, explains Mashile.
Prior to signing up for the 2019 programme, Mashile had completed the first two years of an industrial engineering degree at Stellenbosh University — an opportunity she had to let go of for her the sake of her mental health. When she returned home, Mashile worked as a creative director and stylist for lookbooks and fashion shows to fill the time between applying for bursaries, scholarships and learnerships.
The shift from industrial engineering to data science was informed by her need to merge her interests in numbers with her love for creativity. Through data science Mashile is able to explore a data set with the goal of “telling stories through numbers”.
When the field of data science gained momentum in the early 2010s, its objective was to better the lives of forgettable everyday people by making them visible and thus validating their intersectionalities. The field promised to improve their ways of life because including them in data sets would offer strongholds (ranging from government organisations to large commercial brands like Nike) insights that would result in decision-making that acknowledges the marginalised.
In Mashile’s experience, the field has since developed in a direction that makes data “feel like invisible software that just sits in a cloud that the ordinary person has no access to”. Speaking to Forbes magazine, data visualisation expert Stephen Few reiterates this idea by explaining that although “numbers have an important story to tell, they rely on you to give them a clear and convincing voice”.
Practitioners such as Few and Mashile suggest that large organisations have become apathetic toward infographics, charts and colourful Powerpoint presentations. This is where data storytelling comes to the fore by using data sets to identify more focused and appealing narratives.
According to Mashile, data storytellers are required to ask themselves “who, what, when, where, how and why?” when analysing a dataset in the same way journalists would when writing a news report.
To meet me halfway, Mashile offers the potential use of data storytelling in fashion as an example. Take Gucci. Whether they are authentic or knock-off accessories, the “GC” emblem and the red and green stripes are commonplace in South Africa’s sartorial aesthetic. By finding out where the Gucci buyers are from, their age, job status and how long they have had their Gucci items, such data can help tell the story of the relationship between socioeconomic aspirations and fashion in South Africa.
With it being a relatively new field, data science is still exclusionary. A lot of the stories that practitioners like Mashile are interested in exploring do not have data sets. “That’s where the erasure comes in,” says Mashile. She then adds that data science cannot serve “marginalised groups if it does not acknowledge them”. In response to this, Mashile’s practice looks to one day collect the data of the overlooked.
Before she is prompted Mashile is quick to acknowledge how idealistic this aspiration is. “I could sit here and romanticise data all day but the reality is there’s a lot of pessimism around it,” she said in reference to people feeling that the collection of their data is an invasion of privacy that can often be compromising. An example of this is how Google can record the conversations of people who are near a device. “People need to know that they can trust me with their data because that means they’re trusting me with their stories,” Mashile adds.
As she nears the completion of her training, Mashile is going to Barcelona to take part in an artificial intelligence engineering boot camp after being awarded a partial scholarship. This bootcamp will deepen the sensitivity with which she approaches data science and storytelling because “artificial intelligence replicates the learning process of the human mind by studying large data sets”.
With her departure date set for the end of the month, Mashile is raising money to pay for her tuition and accommodation.
It would be to the detriment of data sciences if the practice remained out of the hands of marginalised folk. For this reason Mashile feels that it is imperative for her to solidify and broaden her skills to increase her chances of being a thought-leader. “We need representatives of what the world looks like because we can never again have a situation where the minority is not considered in the decision making process”.