Data science competitions and supporting platforms Such as Zindi have contributed immensely to the growth of data science and machine learning worldwide. They have created a space where state of the art machine learning algorithms can be tested and put into practice. To businesses, machine learning is an important technology that is going to disrupt many industries. The companies that master it better are going to have a competitive advantage over the others. Data science competitions provide an avenue to not only solve a business problem but also find the best solution with state of the art approaches. To the competitors, these platforms have created avenues where they not only apply their machine learning concepts by solving real-world problems but also an opportunity to learn by getting exposed to state of the art approaches. By showcasing their talents competitors also get exposed to numerous opportunities.
On 26th – 27th October 2019, Ai Kenya gathered over 80 data scientists and tech enthusiasts at iHub Nairobi for the Kenyan edition of the Cape 2 Nairo Datathon. This was a 2 day data science and machine learning hackathon between Nairobi and CapeTown hosted on Zindi focused on the Sendy Ltd Logistics Challenge. Find out more about the challenge here http://bit.ly/sendychallenge.
Here is the BBC World News highlight of the event.
As expressed by a number of participants the competition was not only challenging but it was also educative and fun.The first day started with attendees check-in followed by brief introductory speeches from the event sponsors and unveiling the problem to be solved.The video below highlights some of the speeches from sponsors, a brief introduction to the challenge and feedback from some of the participants at the event.
After the introductory sessions Frida Wanjala, a data scientist at M-Kopa Solar walked the participants through the solution in R programming language. Frida introduced the Cross-Industry Standard Process for Data Mining (CRISP-DM) approach to planning a Machine Learning (ML) project. She then walked participants through the phases involved in CRISP-DM using R programming language. The phases include:
- Business Understanding
- Data Understanding
- Data Preparation
- Modelling
- Deployment
Watch the full R solution walkthrough by Frida here
Frida’s presentation was followed by a second walkthrough in Python programming language by Steven Simba, one of the top data scientists on Zindi. Steven highlighted some of the good practices and tricks to winning a Zindi competition among which he emphasized on:
- Good coding skills
- Good analytical skills
- Knowledge of the best machine learning libraries for the challenge
Watch the full Python solution walkthrough by Steven here
Armed with the skills and knowledge of the challenge the participants broke into groups as some tackled the challenge individually. The challenge continued into the second day until 4pm on 27th November 2019 when it ended and winners were announced and awarded based on the Zindi’s public Leaderboard ranking.
The winners were as follows:
The first position in Kenya was Lawrence Moruye who was also ranked second overall. Lawrence said that his winning approach was mainly based on extensive feature engineering.
The second position was a team from Moringa School consisting of:
- Josephine Wanjiku,
- Agnes Mueni
- Daisy Chebet
- Darius Murori
- Lucille Kaleha.
- Daniel Kuria
- Sharon Nzaya
- Evalyne Wamathi
The third position went to Rodgers Nyangweso
The prizes were awarded as follows:
- 1st Prize : $700 -( Best solution from either South Africa or Kenya)
- 2nd Prize : $400 – (2nd best from either South Africa or Kenya)
- 3rd Prize : $150 – (from each country)
- 4th prize : $100 – (from each country)
The 2019 sponsors
- Sendy: A logistics platform that connects drivers to businesses across East Africa.
- Zindi Africa: A data science competition platform in Africa
- Insight2Impact: Resource center based in South Africa supporting the use of data for decision making with a focus on financial and economic inclusion.
- Ajua: Formerly Msurvey, Ajua is an integrated customer experience in Africa connecting businesses to customers in real-time.
- Cenfri : A non-profit think tank based in Cape Town that collaborates with universities in the region to support financial sector development and financial inclusion by facilitating better regulation and market provision of financial services.
- FinMark Trust: An independent non-profit trust whose purpose is ‘Making financial markets work for the poor, by promoting financial inclusion and regional financial integration’.
- Cape Innovation and technology Initiative (CiTi): a tech incubator that promotes the inclusive growth of the digital economy.
- iHub: An Innovation hub and hacker space for the technology community in Nairobi.
- Explore Data Science Academy: An Ai learning institution and services company based in South Africa
- Whats Good Studios: A leading digital content studio in East Africa
Finally, Ai Kenya wishes to express its gratitude to the sponsors, participants, Data science mentors, Ai Kenya team and the entire Ai Kenya Community for making the event a success. We look forward to hosting more competitions and AI & robotics related events in future. The event photos are available here
Join the Ai Kenya meetup group on https://www.meetup.com/AI-KENYA/ for updates on all our upcoming events and opportunities in machine learning, robotics, data science and data analysis.
Follow us on:
This story was first posted here: https://kenya.ai/cape-to-nairo-datathon-highlight/
