Project Ideas for an Analytics or Data Science Portfolio
If you’re trying to land your first job in analytics or data science and have no relevant experience, the common advice is to build a portfolio of projects to demonstrate your skills.
But where do you start if you have no idea what kind of projects to do?
A few things to keep in mind:
- It’s better to focus on solving real (or realistic) problems and answering useful questions than to do a unique project that doesn’t provide any value. (Real or hypothetical value.)
- It’s better to use the best tool for your project than to try to force a solution with fancy tools that might not be practical or scalable.
- Your projects don’t need to be super unique and original. Many real projects on the job are quite boring and solve seemingly obvious problems.
- You don’t have to stick to one industry for your projects, just pick something that is somewhat interesting to you. It’s better to demonstrate real interest in the subject than force a project for a specific industry just because it’s “hot.”
Projects to solve your own problems
Create a budget. You can do this in Excel or try something interactive using R Shiny or Python + Streamlit. Can you track all of your spending, categorize it, track your progress against monthly budget goals, and see how much is left for the given month or year?
Learn from your job search. Keep track of the jobs you apply for. Not just the company, title, link to the job description, and dates of interviews, but also keep track of the salary range, what percentage of qualifications you have, if you had a referral or not, if you included a cover letter or not, the source where you found the job, how many rounds of interviews you made it through, if the interview loop included a live coding assessment or take-home assignment, etc. Can you start to learn from your own data? What’s the average percentage of qualifications you have for the jobs you actually interview for? How about the average salary? Does including a cover letter or having a referral make a difference in getting an interview?
Download your smartwatch data. Many of us have tiny computers on our wrists capturing tons of data all day, every day. What kind of things can you learn about yourself by analyzing this data? For example, is there a correlation between amount of sleep and activity levels? Do you notice any patterns in your activity — can you predict which days you’ll be active or how many steps you’ll take? Or predict how much you’ll sleep?
Projects to generate recommendations
Figure out your travel plans. Download travel review data from a site like Trip Advisor (here’s a source but check Kaggle and other sites for other datasets). Can you use clustering models to create clusters of similar users or similar destinations? Or use nearest neighbors algorithms to generate recommendations? Can you learn from these and create your next travel itinerary?
Figure out what book to read next. Similar to the above, can you build a recommender system from book rating data to suggest what book you should read next?
Other fun project ideas
Recreate your favorite games. You don’t need something nearly as thorough as an actual online game, but can you write some Python or R code to create a very basic version of Wordle or Mad Libs?
Predict the price of homes in your area. Look for local data — your city, county, state — for recent home sales. Can you create a predictive model to accurately predict a home’s sale price? Try different models, and different features, and see what kind of accuracy you can achieve. Then input a few local homes for sale to see what your model predicts, and keep an eye to see what they eventually sell for.
Need more ideas? Check out this list.
Additionally, if you need to find data, check out this list.
Finally, once you have a project idea, here are tips for how to do it.
Want more career advice? Follow me on TikTok, Instagram, or LinkedIn, and sign up for my free data career newsletter.
