University of Zambia Remotely Hosted Hackathon
DARA Big Data hosted its first remotely held hackathon at the end of September 2020, which was held on campus at the University of Zambia (UNZA). Usually DARA Big Data would host hackathon events throughout the year, however the global COVID19 pandemic has meant that regular events have not been possible and the project has had to find alternative ways to continue to deliver learning.
The 2 day event was put together by DARA Big Data and the IAU Office of Astronomy for Development (OAD), with cloud computing support provided by the Inter-University Institute for Data Intensive Astronomy (IDIA). We were greatly assisted by UNZA Physics academics Dr Habatwa Mweene and Dr Nchimunya Mwiinga, with additional vital support provided by former DARA Big Data student Emmanuel Ngonga and future DARA Project student Jones Chilufya. The students taking part were mainly in the 3rd and 4th years of Computational Physics degrees and were given leave from their regular learning to attend the hack.
UNZA students had returned to campus in Lusaka amid fairly low national COVID19 infection rates, however with international travel still quite risky we had to connect most of the tutors remotely. Luckily Emmanuel and Jones were on hand in Lusaka to act as tutors in the classroom, both of whom had studied at UNZA previously. Dr Brenda Namumba, who is also a former UNZA student and was the first Zambian to gain an Astrophysics PhD, acted as a remote tutor in addition to Dr Madhanpall, Dr Vanessa McBride (from the OAD) and Dr Carolina Odman-Govender (from IDIA). Remote tutors were connected to students via Slack for questions and advice, with tutorials and presentations given using Zoom. COVID19 safety measures were observed on campus and students were all issued with DARA Big Data face masks and hand sanitiser.
The hackathon itself was put together by Dr Nikhita Madhanpall, who works as a DARA Big Data Fellow at the offices of the OAD in Cape Town. Her remit is to develop hacks to be held in the 8 SKA African partner countries to promote data science and machine learning skills across the continent. For this event she put together a fun and highly topical hack which taught students how to collect COVID19-related tweets from around the world and look at them in different contexts. This is known as ‘sentiment analysis’, an automated way of analysing Twitter data to classify it as either positive, negative or neutral. Students learned how to build their own machine learning (ML) models using Python in order to do this and in the process developed some impressive skills.
The first day of the hack opened with a speech from the UNZA Dean of Natural Sciences, who acknowledged the upcoming 4th Industrial Revolution and the importance of learning big data skills. Students were divided into 6 groups with the first day mainly consisting of tutorials and learning how to use Twitter API credentials. The second day saw students setting themselves research questions to interpret their data, build ML models and put together presentations outlining their findings. Presentations were given at the end of the second day and all of the tutors were impressed with the quality of student research and their enthusiasm for the topic and methods. Prizes were given to the 3 teams that tutors felt had put together the best models and the most thorough presentations. The 3rd placed team looked at how people in Zambia felt about schools re-opening, the 2nd placed team tried to assess the impact of COVID19 on the New York economy, and the top placed team put together an excellent analysis on whether people tended to see COVID19 as some sort of hoax.
Student feedback following the event was very positive, we received comments such as 'I can very much see myself as a data scientist or a developer in future' and 'I'm excited to learn more on my own about machine learning and AI and hopefully see where that could take me'. Dr Habatwa Mweene said that the UNZA Department of Physics was ‘grateful that students had an opportunity to see the knowledge from their studies developed and put to use. Students were enthusiastic and hoped that a future face-to-face workshop will happen.’
Former DARA Big Data student Emmanuel said that students ‘went about their hackathon challenge with inspiring zeal, I felt very privileged to work with them and would be thrilled to work with them again. Seeing them relentlessly tackle the hackathon challenge convinced me that there is a place for a vibrant community of data scientists in Zambia. I was especially impressed by the female students who have embraced science in a society that currently has few female scientists.’