A Data Science Conversation
How to teach yourself data science, earning data science jobs through projects, and other lessons from my interview on the Towards Data Science podcast.
Talking is a lot like writing in that it forces you to formulate vague ideas into understandable, concrete concepts. Often, we think we know an idea, but we don’t truly comprehend the theory unless we can communicate it clearly in words to someone else. That forced understanding through communication is partly what drove me to write about data science in the first place, and what drove me to speak recently on the Towards Data Science podcast.
In the episode, we (me and the host YK from CS Dojo) talk about how I self-taught myself data science, how I earned two data science jobs through my projects, how to make learning data science enjoyable, a top-down approach to learning, and other advice from my journey into the data science field. You can listen to the whole episode (and all other episodes of the TDS podcast) for free on Spotify. Below the episode, I’ve captured some of the key takeaways.
- Don’t expect a formal university to teach you material applicable to practicing data science in industry. Online courses, which are updated rapidly and often are developed with companies, are more efficient and cost-effective ways to learn the data science skills you need in industry.
- Data science projects, blog posts, or Github repos can get you in the door for a job interview. Technical and communication skills earn you the job (and continuing success throughout your career).
- Try a top-down approach to learning. Figure out what problems a technology or algorithm can solve before delving into the details of how the method works.
- Focus on building a portfolio, composed of real-world data science projects, to demonstrate your technical skills. Place emphasis on mastering skills instead of acquiring credentials, which don’t mean a lot in the data science industry.
- Make learning data science fun by working on projects that personally interest you or solve a problem you or others around you have. I was motivated to keep learning data science not only because of the prospect it would earn me a good job but also because learning data science was inherently enjoyable.
- Reading, whether technical or for personal learning, is part of a process of becoming less wrong about the world. You come into a book with some background knowledge (priors), learn new information (gather data), and update your beliefs to more closely match reality (a posterior). Try to choose books that teach you something and that you also find intriguing.
The overall experience of my first podcast was positive, and, as someone who has gained a ton from listening to podcasts, it was fascinating to see a podcast from the interviewee’s viewpoint. Kudos to all involved at Towards Data Science (for the blog and the podcast) and I look forward to contributing more to building the data science community.
As always, I welcome feedback and constructive criticism. You can reach out in the responses, or on Twitter @koehrsen_will.
