PSA: Don't Attend Heinz College’s MSPPM-DA Program If You Want to Be a Data Scientist

[hindsightis2022]

For context, I'm a second-year DA student at Heinz who came into this program wanting to be a data scientist. During a fall visitation weekend and through multiple emails with staff during the application process, I asked if it was possible to become a data scientist through this program. I got vague assurances at the time that this was possible. From my experience and the experience of my peers, this program is not worth it if you want to be a data scientist. Please do yourself a favor and look elsewhere, because I really wish that I had.

Here’s a couple of reasons behind my above assertion:

1. There are too many required classes that provide little to no useful value. For instance, if you have less than 3 years of work experience, you’re required to take a public speaking class for 6 weeks, and you are required to take a 6 week course to learn how to write policy memos. Furthermore, you have to take a 6 week course in “organizational design and implementation” where you read HBS cases and decide how you would have approached them. Finally, you have to take 6 units of a “finance” class that is a really basic class where you just learn how to read financial statements. This is not helpful at all for someone who wants to be a data scientist.
2. The rigor of classes here is… questionable. There are a few classes, namely the required Database class for DA students and Big Data and Large-Scale Computing that are actually rigorous, well-taught, and useful. The rest of the classes either provide too surface-level of an approach to the topics discussed or try to tackle a topic that really should have been covered in a semester in 6 weeks, leaving you with little real comprehension of a topic without a lot of outside work on your own.
3. There is a lack of understanding in many classes of real-world tools that should be incorporated in class. For instance, it would make sense, particularly in the programming intensive classes, to mandate the use of GitHub or similar tools for practice with version control. Anyone who is going to program in a real data science job will use GitHub/GitLab and therefore it would be useful to become much more skilled in it through a grad program. This would also help to easily create a portfolio of work to show future employers. Also, (this has changed) but for my class, we were taught machine learning/data mining in R instead of in Python (scikit-learn) which is more often than not the industry standard. There are a few classes taught by professors who have more non-academic experience who help to clue students into the real world tools but this is pretty rare.

Heinz College in general is trying to be too many things at one time and therefore is unsuccessful in being particularly good at one particular thing. The biggest reason peers of mine and myself chose this program over something like UChicago’s MSCAPP program is because of Heinz’s financial aid which is substantial. It was substantial enough for me that I will graduate without any student loans, for context. I am doubting, however, whether the financial savings are worth the lack of educational rigor and course flexibility offered by Heinz. My advice here would be to apply to the MSPPM-DA program for the financial aid package and use it to try to get a better financial aid offer at UChicago (MSCAPP), Georgetown (MSDSPP), or some other comparable program.

I'm happy to answer more questions about this! I really wish someone had been much more direct with me in the application process about who this program is best suited for. I don't want someone else to make the same mistake that I did and leave the Heinz program being as frustrated as I am with my overall experience.

 

[ExponentialDecay]

The question I have is, why did you think that a policy program would prepare you to be a data scientist?

[unbranded]

You missed an opportunity to make your username "Heinzsightis2022"

[Mr. Jameson]

23 minutes ago, ExponentialDecay said:

The question I have is, why did you think that a policy program would prepare you to be a data scientist?

I would say MSCAPP absolutely prepares you to be a data scientist. Strong training in statistics, computing and machine learning, and interpreting data for non-technical audiences. We are fully integrated with the CS department at UChicago. Data science itself is interdisciplinary-- if you are looking to go into software engineering or UI/UX maybe it makes less sense (although I know people doing this from CAPP as well).

[hindsightis2022]

10 hours ago, ExponentialDecay said:

The question I have is, why did you think that a policy program would prepare you to be a data scientist?

As mentioned above in relation to the MSCAPP program at UChicago, the MSPPM-DA program is similarly not billed as a pure policy program. In fact, to date, I've actually only taken 1 pure policy class for the degree. Also, Heinz in general isn't just a policy school, it also is an information systems school, among other things.

I should also clarify that I want to work for non-profits/for the federal government/generally in the "social good" space and harness data science tools to help these organizations achieve their public-service oriented goals. So, when applying to grad school,  I was specifically looking for institutions who do work in the public interest technology space and I specifically didn't  apply to a typical MPP program because I didn't want to just learn about policy/the policymaking process. However, I also didn't want a pure data science program, because these programs typically focus on using data science to solve private sector business problems which I was not interested in. There are programs that are emerging now (MSCAPP, MSDSPP, etc) that seek to blend data science, machine learning, and social good efforts and I was looking for grad programs that would be able to combine these worlds together. Heinz just hasn't quite been what it was billed to be, which is what I'm trying to get across here (perhaps a bit harshly).

 

 

Edited January 13, 2022 by hindsightis2022

[ExponentialDecay]

On 1/12/2022 at 8:15 PM, hindsightis2022 said:

I also didn't want a pure data science program, because these programs typically focus on using data science to solve private sector business problems which I was not interested in. 

I can sympathize with not wanting to spend years learning applications you're not interested in. That was me in every single calculus course. That said, when you're trying to enter a mathematical field (assuming you mean legit data science, with blackjack and hookers matlab and models, not the light rearranging of datasets in stata or excel) and you don't have a mathematical background, that's a major handicap. People who do actual math - in the policy space, in the business space, wherever - tend to be suspicious (in my experience) of candidates coming out of so-called quantitative policy programs unless those candidates demonstrate a mathematical background in some other way (undergrad, work experience). Specifically, because policy programs don't give a strong theoretical component. Like, an econometrics course at a policy program that goes into any significant detail on matrices is a rare thing. If you're doing anything more advanced than plugging stuff into stata, being able to understand the theory behind what you're doing is important (and even then - I've seen people make gross errors with plugging stuff into stata that showed they fundamentally misunderstood the function they were trying to replicate). Fairly or unfairly, someone who comes from a policy program faces a question that someone from an applied math program wouldn't necessarily get.

tl;dr program selection isn't just about what classes you can take, who the profs are, what the placement is, how much aid you're getting and the rest of that wonderfully complex bouquet of factors - it's ultimately about how that program will brand you. And what kind of branding you can benefit from depends on where you want to go, but also on where you're starting from. Policy school is a natural choice for people who want to go into policy (it's in the name, after all!), but in terms of what that brand would do for a specific candidate's background, I feel like a lot of people, especially humanities majors hoping to make the jump to something more quantitative, would be better served by a specialized masters in something like finance, econ, applied math (you can take the prereqs non-degree) or even an MBA (they just have more credibility, even in policy).