New publication showing a causal link between college tuition subsidies and voting
To learn more about working with me on data science tutoring, reach out directly at igorgeyn at gmail dot com—or via LinkedIn/Twitter.
I specialize in taking complex statistics and data science concepts, turning them into concrete and understandable exercises (with or without code), and providing explanations that fit a variety of educational and professional backgrounds.
A BIT ABOUT ME:
I'm a data scientist working at PG&E and a PhD candidate in political methodology and statistics at UCLA. I have published two papers using observational causal inference in top social science journals. I’ve also presented analytical work at academic conferences and in front of large audiences at Fortune 500 companies, published a book chapter in an academic textbook, and have worked on more than a dozen different projects spanning various industry verticals.
Perhaps most importantly: I have been (1) a middle school tutor, (2) a high school tutor, (3) a peer tutor in undergrad at UC San Diego, (4) a teaching assistant (TA) and teaching fellow (TF) at UCLA, and (5) an instructor of record (basically, the main teacher) for courses I developed at UCLA. I’ve taught math and stats at nearly every academic level—including through in-person and digital/remote channels—and bring that wealth of teaching experience to my 1-1 tutoring work.
MY APPROACH:
While I have formal educational training through undergraduate econometrics and graduate-level UCLA courses in the economics, statistics, and political methodology departments, I learned my most important data science skills through applied, hands-on training. I use the philosophy that mastery—especially in data science—is acquired through practical experience and demonstration rather than through lecturing and textbook exams. You flip the classroom.
This means that, in our sessions, we (mostly) won't be:
- spending a bunch of time reading complicated statistical theory or deriving proofs (I did this so you don’t have to)
- doing a bunch of math by hand/with pen and paper
- learning about outdated analysis techniques that are rapidly being replaced by new techniques
And, instead, we will be:
- focusing on what’s important to you/what you want to learn (no one comes in to data science with zero skills/knowledge)
- situating our work in a concrete and specific context that makes sense to YOU (it doesn’t matter if the textbook is interested in sports or the news or whatever, if you aren’t)
- using real data, computer programming/code, visualizations, and interactive tools as much as possible—to SHOW rather than tell you how data science works
Sometimes, we will motivate a new idea by working out proofs and doing statistics by hand—especially if that’s important to you. But we will spend the bulk (~80%) of our time choosing a project that interests you (or choosing from a sandbox example that I have come up with), identifying the key statistical/data science problems in that project, and then solving those key problems with modern programming and analysis techniques (using R, Python, and a visualization tool of your choosing).
WHERE DOES THIS APPROACH COME FROM?
I honed my approach to teaching through two main experiences: assisting with/teaching classes at UCLA and my professional experience as a data scientist.
I had the pleasure of TAing under some of UCLA's best and most innovative stats/data science professors, working with hundreds of students during my PhD and teaching numerous iterations of very difficult courses from intro level statistics to applied public policy analysis courses. This experience taught me to build student confidence and understanding using building blocks—as opposed to the information dump approach—and to quickly sense students' preferred learning styles/approaches.
As an applied data scientist (both at PG&E and in previous roles), I constantly share and explain the work that I'm doing. Often, I'm one of only several people versed in a given statistical technique/analysis, and my work just doesn't get very far if I'm not able to communicate it effectively. Over the years, I have honed my ability to explain complicated analyses clearly and succinctly through dozens of projects involving hundreds (if not thousands) of different stakeholders, each of whom has their own unique background, approach, and line of questions.
I bring this experience to bear on my tutoring work.
WHO IS THIS FOR?
I honestly believe that, today, everyone can benefit from data science skills. For those interested in a data science career, I think this is obvious.
However, I also see DS skills as tremendously useful for those working in or considering marketing or business, those in or considering law or medicine (differential diagnosis, anyone?), people who want a better handle on their personal data/consumptive patterns (health data, spending data, etc.), engineers in basically any discipline, and virtually anyone else interested in going from being a passive consumer of information to actively participating in its interpretation and framing.
It’s a really empowering set of skills.
Some specific groups of folks I think are best positioned to take advantage of developing data science skills:
- High school and some advanced middle school students looking to get a head start on a highly useful skillset.
- College students looking to supplement their in-class instruction or develop a project to position themselves for a job/internship.
- Working professionals looking to beef up their existing skillset with deeper conceptual knowledge in statistics, enhanced programming skills, or guidance on an advanced project in a nuanced field (e.g., causal inference).
WHAT'S THE NEXT STEP?
The next step is to book time for us to discuss your specific goals and expectations.
My approach to working with high school and college students is different from my approach to working with my approach to working with working professionals. It's also important whether you interested in tutoring for a specific class (more emphasis on theory and conceptual knowledge, which can come up in tests) or you are interested in jumping into an applied project and adding something to your resume (more emphasis on applied stats knowledge, programming, and coming up with some work that you can publish to a website and show off to your friends/colleagues).
To get started, just send me a text or email to set up a phone call—I'll take care of the rest!
I look forward to chatting. We can start by talking about whether AI/chatbots are allowed (they are, but in a way you might not expect).
======
SOME FAQs
Why not just use ChatGPT/LLMs to teach myself? The short answer is that you will get a lot of overly generalized, questionably correct information that you won't be able to verify/validate. ChatGPT, Claude, and all the rest of them excel at summarizing—but fall short when it comes to rigorously examining a problem and either (1) providing the correct answer or (2) when unsure, stopping short and saying "I need to look into this." In a nutshell, LLMs are overly confident and, while helpful for spot-checking or performing super basic coding tasks, not great for learning nuanced, fast-evolving subjects like DS.
What are some examples of projects that students have worked on in the past? The main thing I want to convey is that students, genuinely and truly, can work on whatever they want. I come from a social science background so a lot of the folks I've worked with gravitate towards topics in the social sciences (economics, psychology, business, political science, sociology, etc.), but I encourage exploration of any topic in any domain. Just as long as it's interesting to the student.
Past projects include examining the effect of heat/temperature on judicial decisionmaking, the effectiveness of foreign aid on disease outcomes, whether the growth/onset of online sports betting affects student mental health, and the downstream impacts of affirmative action policy on diversity in college applicant pools. Some projects were then presented at undergraduate research symposia/conferences.
If there is anything else you want to know, just reach out!