Training future Data Scientists - Part 4: Growth - it's a process
Sun don't shine in the shade - Kanye West, Waves
Growth is not only for the students, but for the program itself and all those involved. In 2017, growth was definitely on the cards for all involved. It was a time to stretch our methods, our relationships and our processes. Let’s take a look at some of the stretches we made in 2017.
The CSIR is a research organisation. We have many researchers around the CSIR Modelling and Digital Science building. To tap into this talent during the DSIDE program, we work to have constant opportunities for collisions between student teams and researchers. One of these are Friday project progress presentations. These in the past were called Shoe* and Tell (The show was misspelled and we just kept it that way).
We tried to fit in 6 project presentations in 60 minutes. This gave the teams an opportunity to present their work and ideas to other researchers and developers at MDS, while at the same time exposing the rest of MDS to what the DSIDE and Data Science teams were working on. The challenge we quickly identified was that the time was too little (7 minutes for each team with Q&A). In 2017, we altered this model. We introduced both shorter and longer formats of the presentations.
In 2017, we connected a number of times with the Data Science for Social Good program at the University of Chicago and the EU program at Nova Business School in Lisbon, Portugal. One of the things we learnt from these interactions is the use of deep dives to allow the teams to have a long period of time given to each team to discuss their topic in depth which also allows the audience a deeper understanding of the project for better feedback. We took this into our 2017 program with having Deep Dive Fridays. Here we had only 2 teams presenting a week for 3 consecutive weeks, with 30 minutes given to each team. But, before the deep dives the teams has shorter shoe and tells for 2 weeks before that. The Shoe and Tells were reduced to 5 minutes. This was to allow some initial Exploratory Data Analysis to be done and progress be tracked.
The deep dives were indeed deep. What was encouraging was seeing students and researchers participate in asking questions and thinking of the ramifications of decisions being made by each of the teams or team members. When talking to the students, I also realised we had increased the stakes with Deep Dives. The students wanted to make sure that they maximise on the opportunity and also show off as much as they could of how much work they had already done. Highlighted by students was how the deep dives:
- Allowed discussion and collaboration between projects that shared some similar problems (increasing information diffusion),
- Pushed each of the projects forward and encouraged reaching completion
- Encouraged creativity within and across projects, students got to working to share why their projects and solutions were simportant.
For researchers the deep dives acted as sanity checkpoints and also required a good grip on what the direction of the project was. The projects that had partners also benefited from the deep dive in that the teams worked to better crystallise the challenge.
2017/2018 saw us add more opportunity for enrichment outside the lab. While in the past we had some visitors come in haphazardly to visit students or talk briefly. In 2017/2018 we expanded this program. We introduced Friday Townhalls. These are an hour or so sessions where an external speaker comes in and talks about a topic. In this season we had
- Tefo Mohapi, founder iAfrikan (disclaimer, I occasionally write for iAfrikan) - Privacy and Data
- Sipho Nxasane, Innovation, City of Tshwane
- Lydia Molefe, Reserve Bank
- Rapelang Rabana, Chief Digital Officer, BCX
- Anna WIlke, PhD student - Political Science, Columbia University - Data and Political Science
- Prof. Fulufhelo Nelwamondo, Executive Director, CSIR MDS - Role of Data Science in the CSIR
We are eternally thankful for all of our speakers putting aside time to come in. In 2018, we hope to expand this. Get in touch if you would like to participate.
Project team Support
In the past we had more senior researchers work directly with student teams. In 2017/2018 we worked with more researchers to also assign mentors to each team. As such almost all teams had a mentor dedicated with the team who would help guide the team on developing their solution. The mentors are not there to solve the problems of the team, but act as a first line of offence and defense to steer the project. They work with the team, work with the project partner and the senior researchers.
There were many areas of growth, I only covered some notable ones. Below are a few more and I may update them as time goes on.
- We moved DSIDE@MDS to a new space, the recently named Data Commons in the MDS building. After a great suggestion from our group assistant Sylvia, we worked with the Data Science team and support staff to transform the space.
- Tsholofelo, our communications specialist, worked with researchers from other MDS competencies (namely Advanced Mathematical Modelling, Mobile Intelligent Autonomous Systems and Information Security) to get them to give an information session on what MDS does in a broader sense. This assisted in integrating the DSIDE students into MDS day to day functioning.
- Internal presentation day: This year, instead of having the final presentations being the only time that teams really see each other (on a livestream really) we had a big internal presentation day that brought all 50 students (groups from MDS and from Meraka) to present to each other a week before the final exhibition day.
- Having alumni review reports and provide feedback. This was on a volunteer basis and we are forever grateful for the feedback teams got.
- Posters added as an output to the program. In the past only a report and code was an output. This also meant that the teams had posters
This is fourth in a series of blog posts covering the 2017/2018 DSIDE program. In the next few blog posts I'll cover what happens in preplanning, what has changed over the years and where we might go. Thank's to Raesetje Sefala and Thulani Khumalo for inputs on this post.