Skip to content
Site icon Daksh Agrawal
Menu
Back to Blog
Data Science @ Unimelb

Data Science @ Unimelb

Just a fun onboarding pack. Continuously updated (hopefully).

education university careers

Overall Review

It’s great, but keep your expectations in check. Their purpose is to educate you, teach you to think, and show you how the world works. Their purpose is NOT to get you a job.

Academics

Whether you like it or not, and better start liking it, academics will be an important chunk of your life, and you need to take them seriously. Even better if you enjoy it, but not everyone does that, and it’s fine too.

Parallel Pathways

The data science degree is a mix of math and computing, and that’s two parallel pathways you’re travelling together on:

Mathematics: Calculus + Linear Algebra → Probability (for Statistics) → Statistics → Linear Statistical Models → Modern Applied Statistics

Computing: Foundations of Computing → Foundations of Algorithms → Elements of Data Processing → Machine Learning → Applied Data Science

Both pathways are fundamentally important, and you may like one over the other, but you cannot ignore either.

Non-Optional Subjects

As a data scientist, you must know a few subjects that are technically “optional” (due to bureaucratic reasons) but are actually required to do well in your career:

  • Data Analysis: Quick intro to the world of statistics, you can just do statistics, but you will have an easier time if you have done this before
  • Database Systems: Information Modelling, SQL, Query Optimisation, Database Architectures. This is critical for any data science job.
  • Digital Tools & Methods: Too bold of a move to include an arts breadth in here? It’s EODP, but well taught, and your analysis subject is just the arts.
  • Object-Oriented Software Development: Daksh, now you’re just making me do a computing degree. Probably, but gone are the days when you could just chuck a notebook and the ML engineer will figure out the rest of it. Or you might wanna be an ML Engineer, good choice. Take this subject so you know (at least the basics of) how to make your analysis see the light of day.

The Compounding Debt

The worst hellhole I’ve seen students stuck in, including myself. Can’t say about others, but the data science degree. Everything in the third year builds on what you learnt in the second year, and the second year builds on what you learnt in the first year. So if you accumulated debt (half-understood concepts, challenging ideas, missed lectures), you might pass the subject, but it’ll get worse later as you’re destined to struggle with later ideas. A rule of thumb could be that for anything below H2A, you need to revisit the subject over the break and gain an H1-level understanding. Applies to math, applies to computing, applies to life.

How to perform well?

Stay ahead of the curve. Learn things before they’re taught, so university is solidifying concepts, not introducing you to them. Not for everyone, but for those who wanna do exceptionally well.

Subject Recommendations (non-DS)

  • Algorithms and Data Structures: Taught well, good concepts
  • Artificial Intelligence: Not LLMs but still cool things to learn about
  • Information Security & Privacy: Amazingly taught, almost essential knowledge for anyone in tech
  • Computer Systems: Challenging but very helpful
  • Natural Environments: Chill in general

Breadths I took

  • Principles of Management: If you’re interested in corporate management, leadership, entrepreneurship and stuff
  • Managing Processes and Projects: Avoid unless you somehow love operations management and bureaucracy
  • Video Games: Remaking Reality: If you play games, it’s made for you. It’s an arts subject, so marking is generally harsh
  • Logical Methods: Math x Philosophy subject. Amazing for a niche audience, acts as a prelim for MoC.
  • Digital Tools & Methods: See above

Toolkit

A lot of your performance as a data scientist is based on how good your toolset is. You could spend your life in VSCode or Anaconda, but seriously, don’t. There are amazing tools out there, and you should use them for your projects and for efficiency in general.

  • LaTeX: For your reports, projects, and a lot more in general, producing high-quality docs is important, and LaTeX is your friend here. Doesn’t take long, you don’t need to know everything, but try it out. Also, been using Typst recently, and it’s cool too. Here’s a link to my assignment gallery: (Coming soon). Notice how much I did in LaTeX.
  • Deepnote: Hands down the most impressive data science tool I’ve come across. Collaborative notebooks, dashboards, notion integration, whatnot. Amazing data source integrations, too. Impressive shit. Must use for your group projects and anything data science.
  • JetBrains IDEs: Free for students, feature-packed. Maybe too old in the age of cursor.

Crap to Avoid (unless you’re forced to)

  • PowerBI: It’s a bureaucratic hellhole, people might force you to learn it, get certifications and whatnot. But you can learn it within hours and don’t use it unless this is the only BI platform you have access to. Hex, Deepnote, etc can build much, much better dashboards. Put it on your resume anyway, learn it if you get the interview.

Probability vs Probability for Statistics (PFS)

Age-old question. Simple answer. If you love math and proofs, take probebility. Otherwise take Probability for Statistics. Delay that decision until after you’ve done Calculus 2 and Linear Algebra.

Both will leave you unprepared for statistics, both might/might not kill your WAM.

Design of Algorithms vs Algorithms and Data Structures vs Nothing

It’s an elective, so you have the choice of nothing too. If you loved Foundations of Computing and Foundations of Algorithms, you should do a second year algorithms subject, otherwise skip. If you love math and proofs, do Design of Algorithms otherwise Algorithms and Data Structures will give you a generally easier life.

Part-time Work

Melbourne is costly, managing your expenses can get hard. Part-time work is an easy way out, but don’t mindlessly start working in a restaurant, supermarket or doordash at minimum wage, because that’s what everyone does. Nothing wrong with that but there’s better options:

  • Tutoring: If you scored well on VCE or any university subject, you should apply to tutor that, university or third-party. I tutored Foundations of Computing and Foundations of Algorithms at University. Pays well, respectable, and fun!
  • Startups: If you can network well, a lot of young start-ups are willing to hire to part-time devs on random projects. Pay might be shit, but you’ll get perks, learn a lot and have flexibility. Don’t have to believe in the idea, as you aren’t committing.
  • Freelance: Again, networking dependent and a lot more skills needed in general, but amazing work and pay.

Not possible for everyone, but I’d recommend spending your first year purely upskilling, forget part-time, so you’re able to land good jobs from year 2, rather than grinding out tiny money in exchange for long hours, and learning little. You don’t need to get a role in DS or Computing, but anything else you’re good at works too. Passionate about fashion? Apply at a boutique. Love boxing? Train some kids. Melbourne is huge, you’ll find your gig, and make sure it keeps you happy not just pays the bills.

Social Life

Not the best guy to advice on this, I’m one of the least social out there. Still, highly recommend going to CISSA and DSCubed events, and even joining the committee. Absolutely do not miss out on hackathons and any other competitions. It’s not about winning, its building something resume worthy, socialising, learning and having fun. Food too.

Careers

It’s hard out there, bud. You cannot just complete your degree and get a job, doesn’t quite work like that anymore, at least for international students. Build projects, volunteer, compete, show off, blatantly. You need outreach, network and skills above the average to land a role. How to do it well, that’s for another day, but please don’t just sit back and wait for things to happen to you.

YOUR DEGREE IS NOT COMPLETE. You need a lot more skills, lot more experience, lot more projects to be a real data scientist. University taught you a lot, but not all, it’s your job to learn it. Some topics I feel you should learn as an aspiring DS professional: Network analysis, Scraping, LLMs, HQ Visualisation, Storytelling, Cloud, Web Dev, MLOps, Git, Security, APIs. In addition, you need to be thorough with the industry you are working in. Like for TasWater, I had to learn about supply chain, procurement, approval processes and for Square Peg, a lot of finance.

FAQs

  • Does WAM matter? Does anything matter? Boom existential crisis. Dean’s list is a real flex on your resume. H1 WAM is a real flex on your resume. Quant trading will only hire H1 flexers. A fail will lose you time and money (same thing either ways).
  • Should I skip Foundations of Computing? Why would you? If you’re good at programming already, take it as a WAM booster in the summer. If you’re not good at programming, do it since you’ll learn it in a nice structured way rather than cramming ideas up to pass the test. They even do advanced tutorials now, so you can actually learn some content too, all while breezing through your projects and exam.

And don’t forget to enjoy!