Finding the Question

ChiPy Mentorship Blog 1 of 3

February 15, 2018

Meaningful analysis begins with asking the right question for the data at hand. But this process encompasses an art form in which I am an amateur. It has been 19 days since my acceptance to the ChiPy Mentorship, I have sniff at what my ultimate project will be, and I am diving head first. In this post I walk you through my learning process and hesitation, the resources and mindset that carry me, and the project bubbling to the surface.

My Place in the Data-verse

I sit as Impact Analyst at City Year in Chicago, one of 28 cities in which City Year provides student success coaching for populations most at risk of dropping out of school. These student success coaches are AmeriCorps members -- full time volunteers who donate roughly a year of their time in support of a worthy cause. City Year is the largest AmeriCorps program, enlisting more than 3,000 AmeriCorps members (ACMs) per year.

City Year collects a lot of data. Our school districts collect more, and our cities more still. Yet as a nation we have not solved the problem of high school drop-outs and equitable access to quality education. I won't solve it either, but maybe I can affect progress on a small facet of the problem.

There are not many people in my organization, especially not in Chicago, who are familiar with programming, algorithms, and machine learning. This opens up fantastic opportunities to learn and have impact. For example, I wrote a simulated annealing approach to our process of assigning ACMs to school teams based on various features of corps members and schools. This saved the average ACM 90 hours of commuting compared to the previous year.

It makes sense to me that my project should exist in this space of opportunity and impact, but that still leaves a large expanse of potential projects.

Prioritize Learning, Results will Follow

The last two weeks have been an overwhelming and exciting process of brainstorming with my rockstar mentor, Aly. In the process I reached out far and wide to my manager, siblings, the VP of our Analytics department, and the mentors throughout my career. Although this period opened my eyes to a world of possibilities, it simultaneously induced anxiety and debilitation.

I am ambitious and impatient. Analytics excites me because it can have profound effects on problems people didn't know existed or didn't know how to solve. In turn, I was desperate to choose an analysis question that would neatly lead to impact. I worried I might go down a path only to realize it wasn't going anywhere, that the analysis was not actionable, that in the meantime somebody else solved the more fun or obvious question, or that my approach would become obsolete in the toolbox of my organization.

This anxiety was resolved through mentor support and reflection on trends in my own development. Of course results are important, and one should place an educated bet on their approach. But results are not the ultimate goal in this journey. After all, I cannot predict the future. There are uncertainties that may undo any endeavor. More important is the learning process -- being challenged and providing evidence of a strong effort.

Immersion Learning and Mentorship

With personal development squarely as the top priority, I structured my environment accordingly. I typically spend commuting and relaxation time listening to podcasts, scrolling Reddit, and reading. So I infused these spaces with information that reinforces the concepts I am learning. I swapped out my reddit account with subreddits like dataisbeautiful, datascience, learnpython, Python, and Programming. When browsing for a podcast, I look through Talk Python to Me and Python Bytes.

I also went back to basics. Though I have regularly utilized Python through packages like Pandas, Selenium, and Numpy for about a year, my mentor urged me to work through the official Python Tutorial. In just the first 3 sections I had learned several helpful tidbits I had not seen elsewhere. My mentor also shared an article, Teach Yourself Programming in Ten Years. The piece explains how shallow learning can be quick and easy, but a solid foundation and mastery require persistent practice and patience. "A little learning is a dangerous thing," said Alexander Pope.

The common thread in these pieces supporting my mindset is my mentor, Aly. You can check out his blog here. Mentors in ChiPy are uncompensated and contribute 5-10 hours per week supporting their mentees. Mentorship is an incredible resource to any learning effort and a very selfless thing to give. I am grateful for the ChiPy community and Aly for providing that. I hope to provide something valuable in return.

Project Proposal

I do have a project taking shape in my head. I am interested in the variables within and between AmeriCorps members that make their City Year experience successful or not. Can we predict who will need which types of supports? What relationships exist between the training we provide and the services that are delivered to students? Can we better differentiate our feedback and training to provide more appropriate opportunities for growth? Can we improve marketing and hiring practices? Can we create better teams?

The relevant data are diverse, a bit messy, and sensitive at the student level:

  • Notes from observations of ACM tutoring
  • Details about each tutoring session (date, date entered, duration, skill emphasis)
  • Student outcomes (assessments, attendance)
  • Notes from ACM performance reviews
  • Job application essays and notes
  • ACM surveys
  • Teacher and administrator surveys
  • Training schedule and ACM attendance
  • Training feedback from ACMs

To comment on this post or my project, please reach out through the contact button at the top.