Data Associate Program (DAP) 2019
Wednesdays. The middle of the week, humorously also known as “hump day” (Go Google it, it’s not just a me thing. haha.)
Wednesdays. The peak of my week. Emerging from 6 hours of back-to-back classes, meet-ups/calls with project teams and external partners, I would typically like to call it a day.
But Wednesdays are special!
Wednesdays are when we have our weekly Data Associate sharing sessions. And that makes Wednesdays the highlight of my week.
But before diving deeper into these sharing sessions, maybe I should take a step back and revisit the Why and How of the DAP…
Data Associate Program — Vision and History
The Data Associate Program (DAP) is a one of the key long-term programmes of our SMU Business Intelligence and Analytics Club (SMUBIA). SMUBIA, or known in our university simply as BIA, is a student-led club in Singapore Management University (SMU) that focuses on connecting students, alumni and institutions who are passionate about data science.
DAP 2019 was set-up with a vision in mind:
To provide a safe and open community where we can engage in active-learning in the field of analytics and data science.
Breaking down this vision, there were a few key important pointers that we took away:
… a safe and open community…
- Fundamentally, the individuals that make up the community must feel that they belong.
- And on top of that, for a community to be sustainable, it must be driven by a common interest, as well as offering growth opportunities for every one within.
… engage in active learning …
- The Data Associate Program is founded upon a combined shared interest in analytics and all things “data”. We should be a fruitful community where everyone is motivated to pursue their own varied interests in this wide and exciting field.
A couple weeks passed as we pondered about the “How?” of the DAP program, about how we could build a community that fulfils these lofty goals that we had in mind.
Common interest? Check. Data Science.
Community? Sustainable? Motivational? Growth-oriented? Eh, now this is hard.
Last year, the core team conducting weekly study groups for the associates. But that devolved into watching youtube videos, reading articles and following guides/walkthroughs. All particularly non-community-like activities.
The core team then attempted to engage the associates with direct lecture-style sessions, but this was very draining for the core team. Not sustainable at all.
How best, then, can we build a sustainable community which intrinsically motivates all members to grow?
And then the “Aha!” moment came.
Project-oriented Learning
What if every associate could determine their own interests and let those interests guide themselves in their learning journey? After all, each associates has a different set of interests and goals, and a proper curriculum would never be able to align with everyone’s individual goals.
share ideas and celebrate the little successes
What better way, then, to have the associates learn through doing. Working on self-initiated projects in small groups, projects that they themselves would own and give life to.
Thus, the first part of the Data Associate Program, self-initiated DAP projects, was born.
Our weekly sessions were thus formed to be a platform in which the different DAP project groups could check in on each other, share ideas and celebrate the little successes along each associate’s journey.
Book Chapter Sharing Sessions
Really proud of this project-driven approach, we bounced this idea off some of our industry contacts. Most of them were pretty receptive of the direction the program was taking, but they all had one major concern regard this: this mode of learning cannot replace a formal study of major topics prevalent in data science.
Inspired by the how many post-graduate programs and research centers conduct regular reading groups, we decided that DAP 2019 should have a similar arrangement introduced as well. Each week, a different group will be in charge of leading a sharing/discussion about a particular topic. We toyed around with self-selected topics, but finally decided on a more structured approach.
Many of our industry contacts have suggested various essential textbooks, in particular Gareth James’ Introduction to Statistical Learning (a.k.a. ISLR). We thus subsequently decided to adopt this book as the main reference material for our weekly sharing sessions.
Chapter Sharing Session #1: Linear Regression
Last week, we had our first DAP get-together session, and we announced the teams that would be allocated to share the various topics from ISLR. The first team, sharing about Linear Regression (Chapter 3 of ISLR), consisted on Frans, Muskaan and Sherman.
Theirs was a tough job to accomplish. Firstly, they had extremely short notice (only 1 week) to read up on the chapter, digest the knowledge points and share it with the rest of the associates. On top of that, Sherman, a freshman in SMU, hasn’t even completed his Introductory statistics course!
The pressure was up for them, and time was ticking. Muskaan, a junior from the School of Economics, was more comfortable with the topic at hand and was able to coach Sherman and Frans to understand the key knowledge points. However, digesting ISLR was still going to be a challenge for all 3 of them.
A week later, and the results were astounding.
The team conducted a complete and comprehensive sharing session of Linear Regression (LR), beginning from the foundations with Simple LR, discussing the intuition behind minimising sum squared errors (Hint: Major knowledge point for many predictive models), and extended the discussion to include multivariable LR, dummy variables and interaction terms. The team also discussed the problems of using LR (Residual plots for the win!) and shared key algorithms used for variable selection.
The team even demonstrated how to implement LR using Python!
I am sure that all our associates, and even the mentors, had many take-aways that Wednesday evening.
In particular, shadowing their journey through this one crazy week was a very humbling experience: to see Muskaan stepping up to coach and mentor Frans and Sherman, both of which rose up to the challenge and worked extra hard to absorb the content.
Taking a step back, to reflect on the past week, I am very grateful to have witnessed this mini- learning journey of the Linear Regression Trio.
And penning my thoughts down now at the start of DAP 2019, I am really glad to see Vision DAP 2019 becoming true.
Vision DAP 2019:
To provide a safe and open community where we can engage in active-learning in the field of analytics and data science.
Here’s to more Wednesdays ahead!
