In partnership with
Your Titanic Survival Prediction project isn't impressing anyone. Neither is the Iris Flower Classification. Or the Boston Housing Price Prediction.
Delete them.
I've reviewed 50+ portfolios. Every bootcamp graduate has the same three projects. They look identical. They prove exactly one thing: You know how to copy-paste code.
If you didn't spend 80% of your time cursing at the data because of bad formatting, it’s not a project. It’s a tutorial.
Kaggle datasets are a lie. They are pre-cleaned. The columns are named sensibly. The data types are correct. Real data is ugly. It has emojis in the date column. It has merged cells. It has NULL values where the revenue should be.
What recruiters Actually See
When I open a GitHub repo with "Titanic ML Project" in the title, I close the tab.
But if I did look, I know what I’d find: train_test_split, a Random Forest model, and an accuracy score. What does this tell me? That you can pip install sklearn. That is not a hireable skill.
In the real world, stakeholders don't hand you clean CSVs. They send broken Excel files. They ask, "Why are sales down?" and give you access to a database with 47 tables and zero documentation.
I don't hire code. I hire curiosity.
3 Projects That Actually Get Interviews
1. The Personal Audit Download your own data. Swiggy, Uber, Spotify, or your Bank Statement. This data is a mess. It needs cleaning. But more importantly, it tells a story.
Boring: "I predicted Titanic survival with 82% accuracy."
Hired: "I analyzed my Swiggy history and realized I order 40% more Biryani when I have deadlines on a Friday."
You aren't just another candidate analyzing public data. You are analyst who solves problems.
before reading on. A quick shoutout to our sponsor who keeps this newsletter free. It helps us keep the lights on if you check them out. (Continue Reading)
Introducing the first AI-native CRM
Connect your email, and you’ll instantly get a CRM with enriched customer insights and a platform that grows with your business.
With AI at the core, Attio lets you:
Prospect and route leads with research agents
Get real-time insights during customer calls
Build powerful automations for your complex workflows
Join industry leaders like Granola, Taskrabbit, Flatfile and more.
2. The Scraper Kaggle is a restaurant. Real life is a farm. Show me you can hunt. Scrape a local real estate site. Pull product reviews from Flipkart. Gather job postings from LinkedIn. Every analyst I've hired eventually had to build a dataset that didn't exist. Showing you can scrape data puts you miles ahead of the "Pandas Read CSV" crowd. Plus, you’ll learn about rate limiting and HTML parsing, skills that actually show up on the job.
3. The Boring Business Fix Find a local business. A gym, a shop, or a restaurant. Analyze their Google Reviews. Fix their messy inventory spreadsheet. The project sounds boring? Good. Real analyst work is boring. We aren't building neural networks every day. We are figuring out why revenue dropped 3% last quarter. If you can say, "I found 3 menu items that are wasting 15% of your inventory," you beat the candidate with the "94% Precision Model" every single time.
Stop Hiding Behind Code
Your README.md is more important than your Python script. Don't just paste charts. Answer the "So What?" in your github project:
What problem did I solve?
What was weird? (If the data confirmed what you already thought, it’s useless).
What should we DO? (Analysis without action is just arithmetic).
A simple bar chart that saves money > A complex Neural Network that does nothing.
Delete the Titanic. Build something real.
P.S. If you want to put this into practice, I just added 10 domain specific free projects here https://realanalystjobs.com/projects.