Africa needs data scientists: What will it take to train them?

What makes a good data science and how can African countries fill the shortage?

The lack of skilled labour in the data science and software engineering sectors is a problem worldwide but it is especially true in Africa.

According to Dramane Traore, the Founder of Data Fintech, a Nairobi based data broker, the need for data scientists will shift when businesses realise the value in evaluating their own data and using public domain data.

“Employers, (private and public) and especially top management within those organisations are behind the curve when it comes to use data to make strategic decision. As a result, the market is not creating the conditions for more data analysts,” says Traore.

As bootcamps gain popularity we look at their value in comparison to a computer science degree. Which is better? A coding bootcamp or a computer science degree?

Government bodies and organisations everywhere are digitising their business and this is producing huge amounts of data but Africa has not been progressive when it comes to mining.

“The amount of data will keep on increasing even more rapidly,” Irene Wanjugu a technical mentor at Moringa School, a Nairobi-based coding school, tells IDG Connect.

“We have seen that how well this data is ‘handled’ determines the survival or fall of economies and companies. This is also the case for African economies. The field of data analysis is therefore of paramount importance to the advancement of these economies and systems in Africa.”


What makes a good data scientist?

Wanjugu classifies the qualities needed to get into the data science field as two-fold. One, is getting a grip of the tools for data analysis. She terms this as the hard skills. The plethora of tools that are out there can be overwhelming but they are becoming accessible due to the expanding nature of internet connectivity.

The other set of skills are the soft skills, Wanjugu said. These are the non-technical ones but can lead to improved delivery of analysis.

“A data analyst needs to want to dig deeper into the data. Not just accept the initial results provided by the data but to ask why the results are as they are. Digging deeper is a mind-set,” says Wanjugu. “They need to think outside the box – both in the methods they use for analysis and also when displaying and reporting the findings of the analysis. They should be able to adjust the findings on the fly, regardless of pre-conceived notions about the data.”

She adds that the art of storytelling and problem solving are key in presenting the findings of any analysis.

“A data analyst must combine scientific skills, such as statistics, mathematics and communication skills,” said Traore. “A good data analyst is a professional able to understand business issues and, using data analytics, provide actionable recommendations to the management.”

It's all about helping decision makers to navigate around tones of data, and highlight what is really important for the business.


How can African institutions train good data scientists?

“The term ‘Data Scientist’ is essentially a fancy word for an experienced statistician; that is, a statistician with a modern twist,” Wanjugu explains.

She adds that institutions in Africa have been teaching statistics and mathematics for a very long time and it is about time they combined this with the tools that are out there in the market place.

The language R and statistics have been the traditional skill set data scientists need to have. Most of these skills have been taught in the confines of a traditional education setting. However this is changing. Many programmers are accessing their skills online and with improved connectivity, African students are not left behind.

Python is another growing toolset in the data science field. Pandas in particular, is widely used for simple to complex data analysis. Data visualisation tools such as, ggplot, bokeh, matplotlib, seaborn, pandas (visualisation) and many more libraries are becoming popular. Having a good handle on these tools is a catapult in gaining valuable skills.

In addition, there are several areas that aspiring data scientist can specialise in. This would include general programming skills, data mining, data cleaning and analysis, and data visualisation all which fall under the data analysis chain.

“I would suggest that institutions include more practical assignments to their students. Let the students identify a problem in their society and solve it using what they have learnt in university. Make them think. Make them use their brains. Then they will learn how to apply data analysis techniques regardless of company or industry,” Wanjugu says.

For Traore, training great data analyst touches on education policy, business practices, and infrastructure.

“From a strict business perspective, a quick fix can be the creation of big data consortiums among African top businesses, in collaboration with the sector pure players. The size and complexity of the projects will require large teams of data analysts, paving the way for additional job creations and development of local skills,” he explains.

Organisations also can be on the forefront in creating these opportunities by opening up their data for analysis and also training in house marketers and sales people in using data. It could be as simple as using data tools such as Excel to evaluate sales and leads, to producing reports using Tableau.

In this line, Moringa School is looking to launch a data analyst course for the Kenyan market. Data Fintech on the other hand is using its in house resources to bridge this gap.

“As a pure player in the Big Data Industry, we participate a lot in creating a data analyst workforce by recruiting and training them. Today, 70% of Data Fintech's team members are data analysts and data scientists,” Traore says.

Currently Data Fintech conducts a recruitment program dubbed, ‘Data Fintech Draft Day’, to create a reservoir of data scientist in Nairobi that can be contracted on-demand.

All in all, the maturity of the education system and the business environment in mining data could propel this field and create many skilled data engineers.