Data Science and Machine Learning in Practice — The Nairobi Way
One meetup I recommend in Nairobi for tech people is the Nairobi Women in Machine Learning & Data Science.
First of all, this meetup is not just for women, so guys, you’re allowed to attend. The community has been growing with each meetup and it’s been a source of great value in terms of connections as well as knowledge gained.
The theme of the meetup on Saturday Nov. 25, 2017 was Data Science and Machine Learning in Practice. It kicked off with Anthony Otieno who gave an example of how his team collected data from supermarkets to quantitatively analyze consumer behavior about common number of items purchased, most common shopping time, etc.
Anthony admitted that the thing that took most time was getting the actual receipts because they had to do so unofficially by collecting them from the trash cans, which says something about their level of self-motivation. However, it does introduce a level of bias as the receipts they were most likely to collect were those that consumers were most likely to throw out, which was why — not surprisingly — a lot of the receipts had less than 3 items purchased.
However, there are many possibilities for where such a project could go and if you have access to supermarket data in Kenya and would love to reach out to the team, you can reach them on twitter
and read the complete report here; https://blackorwa.com/2017/11/08/kenya-supermarket-report/Another interesting talk was about Habari Kibra which was introduced by co-founder Michelle Mulemi. Habari Kibra trains Kibera slum teens in analytical journalism, which is a branch of journalism that uses data to answer complex questions.
I particularly loved this idea for many reasons;
- It gives disadvantaged slum kids the opportunity to acquire practical skills for the job market
- The focus is on issues such Healthcare, Education not just politics.
- It’s for-profit and not an NGO
The project only started in February 2017 and volunteers are welcome to contact Michelle on twitter (@michellemercy).
Another presentation was given by Stella Ngugi about Jobonics, which is going to use machine learning to provide recruiting solution for employers to reach, engage and hire top talent. With plenty of experience in HR, Stella is building this to solve problems she used to encounter in her job, such as personal bias or filtering out applications with non-relevant experience…
Catherine Gitau, who formerly interned at Brave Venture Labs, was brave enough, pun intended, to give us a quick walk through of their data use case despite not being on the agenda. Brave aims to provide algorithmic HR, much like Jobonics, except they leverage social media. Through sites like GitHub, they gather data on individuals skills and competencies and then match them to opportunities they may be a good fit for. This fresh take to HR should lead to better fits for roles as technicalities which sometimes lock out good talent take a back seat and things that matter, like how well your code is written, come to the forefront of the process
The interesting thing about this while I admit Jobonics is an amazing idea for the traditional job marketplace, I don’t think I’m going to use myself for the time being. As a person who’s pivoting into a new career outside of academia related to Digital Rock Physics to Education Technology, I’m sure my CV will be one of those that are continuously filtered out. It would be interesting to see how it does compared to popular job-posting platforms Fuzu and Brighter Monday.
