- Therveshan Moodley
Deep State - Data science and machine learning in South Africa
Kaggle recently hosted their third annual Machine Learning and Data Science Survey. The intention was to have a comprehensive view of the state of Data Science and Machine learning around the world. The survey was live for three weeks in October and a full description of the questions asked and the format of the survey can be found here : https://www.kaggle.com/c/kaggle-survey-2019/overview
The results data has been made available to the public and we thought it might be interesting to have a look at some of the South African responses, and how they compare to the rest of the globe.
While there are only 88 South African data points (i.e. employed professionals), it is a starting point that we can hope to look back on as future surveys are run.
There are certainly many questions that can be asked of the data and Kaggle has opened up their own competition with prize money for those who are able to generate the most valuable insights.
In looking at this data, we asked very simple questions as a starting point and the initial direction of our insights seems to confirm that in terms of numbers, South African data professionals are almost in line with global averages. The slight difference seems to be that in South Africa, there are marginally more mid-range data professionals as compared to the very junior and more senior professionals found elsewhere.
South Africa also has a higher percentage of experienced coders than the global average. Furthermore, a higher proportion of South African data professionals are skilled in every tool category.
There is still a massive gender bias across the board as 83% of the participants were male, again in line with the global ratios.
With its roots in Africa, Zindi is a 7,000 strong community of data scientists and organisations who collaborate to solve urgent, data-related challenges in a range of disciplines across healthcare, development, logistics and financial services to name a few. Contact Zindi here if you'd like to learn more.