The first time Sharla Gelfand gave a guest talk at the Toronto Data workshop, there were about a dozen people present in the Bissell Building for a brown bag lunch session. A freelance developer of R, a programming language for statistical computing, Gelfand showed a small but attentive audience how to clean up public data from the recent federal election and then did some quick analysis. That was back in October 2019 before anyone had heard of COVID-19.
Fast forward to August 2020 and Gelfand is back by popular demand. This time, Gelfand’s on Zoom sharing insights about how to enable easy access to data and to replace manual, redundant processes with ones that are automated, reproducible and repeatable. There are four times as many participants in the audience, many of whom are a long way from Toronto.
The Toronto Data Workshop — which is organized Faculty of Information Professors Rohan Alexander and Kelly Lyons and Master of Information student, Faria Khandaker — has proven something of an accidental beneficiary of the pandemic. Not only is attendance up in the group’s growing community of data scientists from academia and industry, but the workshop is also attracting speakers and guests from Montreal, Warsaw and Berkeley, California among other locations.
“Early on, I turned down people who asked if they could attend remotely. I thought there was something to be said for coming in person,” says Alexander. “It turns out I was wrong.”
To keep the sense of community, the workshops, which are held on Thursdays from 4 to 5 p.m., all get underway with three-minute breakout rooms. Participants introduce themselves and answer an icebreaker question. Later, after the presentation and formal Q and A session have wrapped up, those who want to can stick around for informal chatter.
The topics on the agenda have also evolved along with the workshop, which was originally focused on data cleaning. The workshop now looks at all the initial stages in a typical data science workflow, which Alexander says tend not to be “emphasized, taught and talked about as much.” These include data gathering and scraping, data cleaning and preparation, and data sharing and dissemination, as opposed to the more highly discussed later stages such as data analysis and data communication and visualization.
Over the summer, workshop guests have discussed topics such as estimating global maternal mortality, Tokyo AirBnB datasets, and how to link data from the 1940 U.S. Census with the country’s Social Security Administration mortality records.
This summer, in response to protests highlighting the extent of the problems faced by visible minority groups, Alexander and Lyons also announced a new research stipend for visible minority students or recent graduates in Information and Statistical Sciences. Three recipients were awarded stipends of $1,000 each to prepare a research paper on a topic of their choice under the supervision of a faculty mentor. They will later have the opportunity to present their paper at a meeting of the Toronto Data Workshop that suits their research timetable.
Recipient Erwin-Jeffrey Komguem, who graduated last spring with a BSc in Statistics and a double minor in math and computer science, is using the stipend to co-write a research paper exploring the effects of traditional and inverted teaching methods on the learning experiences of students in first year calculus.
Inverted teaching involves preparations like watching a pre-recorded lecture and reading up in advance of class time, which is then used to ask questions and solve problems. Traditional learning means getting the relevant information imparted in class and then solving assigned problems independently.
Together with Professor Alfonso Gracia-Saz of Mathematics, Komguem is analyzing data collected from 2016 to 2019 capturing student grades, performance, course selection, program selection and engagement, among other measures. “I wanted to participate in a research project and put all the skills I had learned at school to use,” says Komguem, who is now working as a data analyst and thinking about doing a Master’s in Statistics or Data Science. “(The stipend) gave me the opportunity to really immerse myself into the research world. I’ve been able to learn a lot. Writing a paper like this is not something I’d done before.”
The stipends are being funded by the Faculty of Information, the Department of Statistical Sciences where Alexander also teaches, and Lyons’ and Alexander’s personal research funds.
“We thought that establishing these stipends was a tangible start and an expression of our support,” said Lyons. “Having been the recipients of small stipends ourselves in the past, we knew that it could make a difference, and that it was something within our power to quickly establish.”
Added Radu Craiu, Chair of the department of Statistical Sciences, “We are always happy to enlarge the Data Science community and we continue to seek new transformative ways to break down barriers.”
You can find information on upcoming workshops here
