If you’ve been on my list for awhile, you probably know that I feel that personal projects are a must for getting a job as a data scientist if you have no professional experience.
Just doing coursework isn’t enough because everyone has done the same courses that you have.
I could go on about the importance of doing a project, but I want to share something with you instead.
I’ve been talking with beginning data scientists like you recently about projects. And it sounds like a lot of you see projects as a necessary evil. That is, you know you have to do some, but you’d rather not. And part of the problem is coming up with a good project in the first place.
One person told me, “I wish someone would just give me a project.”
Rest assured, you’ve got an entire lifetime of people giving you projects, whether you like them or not.
When you get a job, you’ll work on what they tell you, at a level of fidelity they dictate, and in a timeframe that you probably wish were different.
I encourage you to shift your mindset around this and pick something you actually will enjoy working on.
The most important thing about doing a project is…doing a project.
A lot of people get concerned about doing “the right” project. Yes, some projects are better than others. And if you know which field you want to go into (e.g., marketing), it does make sense to do a project in that area.
But if that’s not the case with you, then pick something you would find interesting. Because looking at it as a chore is going to make it tough to slog through.
Great project ideas are all around you! I once saw a project where someone used webscraping to extract results from an annual running race and analyzed the performance of the runners over time. A month ago, I told you about a project from one of the Break Into Data Science members who used text mining to categorize holiday recipes.
One of the more inventive ones I’ve seen is someone who analyzed famous literature to determine a “fingerprint” (via text mining and clustering) of the author. Each author has their own style and this person was able to show, graphically, how Shakespear’s style differed from Hemmingway’s. You could use this type of analytical writing style analysis to determine if a student had copied their term paper off the web (e.g., see if the paper had the same “fingerprint” as the rest of their writings).
The following project—done by one of our Break Into Data Science members—is a beautiful example of doing it right. She used text mining to categorize cooking recipes! It’s an advanced example so don’t worry if your level of data science isn’t as high as hers, but this features:
+ Choosing a really cool project
+ Obtaining data (she used web scraping)
+ Providing the right level of detail
+ Great visuals
+ Providing thoughts on future work and improvements
+ Using a blog post (in this case, GitHub.io) to showcase one’s work
You’ve got to check this out: https://hengrumay.github.io/MenuPlannerHelper/
What, besides data science, are you interested in? Sports? History? Psychology? The environment?
There are scores of freely available data sets out there. As you walk around your daily life and find yourself wondering about something, write it down! Then see if there is a problem there you can analyze.
Take advantage of this time to do something you will really enjoy. Not only will it make the effort more enjoyable, you’ll be much more enthusiastic talking about it with prospective employers.
To your success,