My power is out today. Having no wifi makes it difficult to prep for teaching my next class, so with all of this free time I decided to write my eagerly-anticipated third blog post.

Last year, when I went back into academia, I started a podcast which I called Data Science Beginnings. Doing a podcast is not something I ever thought I wanted to do, I’m not even a fan of podcasts myself, but I thought it would be an interesting way to meet people, provide a service to my department (which is good for earning tenure), and fun to learn how to interview people and put episodes together. All of these have turned out to be true – the tenure part is TBD.

Once I decided to do a podcast I needed to make a bunch of decisions, more or less in the following order.

Who is the audience?

I decided quickly that my audience would be people who are starting their journey into the data sciences, or are thinking about it, i.e. current or potential future students. This seemed to make sense to me given my role as an Assistant Professor in a DS department.

What kind of a show is this? Interview? Solo? Panel discussions?

Our department has a regular seminar series where we invite folks working in industry to give a talk. I thought this would be a great way to find people to interview – just ask the people who are already coming to campus anyways. So I decided to do an interview show. It has not been as easy as I thought it would be to find people willing to be interviewed, and so my dream of having a pipeline of already-scheduled interviews never came true. Instead, I started off by interviewing friends. And then current faculty in my department. And then former students. And probably next will be current students.

What kinds of questions should I ask?

Being a show for people who are not yet working in the data science field, I figured I’d ask questions about how other people got started in the field, how to learn and keep up with data science, and how to get a job. I did not think it appropriate to start discussing very advanced deep learning methods, or reviewing the latest publication. Wrong audience for that.

How often should I post episodes?

Given that this is my first year teaching in the program, I don’t have a ton of time to be scheduling, recording, and editing episodes, so my postings have been sporadic. I think I’m averaging one episode every 1-2 months, which is not great. But, I’d like to post something approximately once every two-three weeks – that’s my goal. And not posting anything during the Summer, cause that’s my time off.

How do I actually record, edit and post episodes?

I was able to convince my department to purchase a podcast recorder, two mics with mic stands, and two headsets. I record episodes in-person in a fairly quiet room on campus. I experimented with recording through Zoom, and will probably not do this again. I edit the episodes in GarageBand. I got an account on Anchor, which is now called Spotify for Podcasters, and post episodes there for free. I created a title for the show, a logo (created in R using a version of this code: I grabbed some license-free music and modified it with some lyrics. I even created a Twitter profile for the show which I have yet to use.

Is what I'm doing interesting enough, or am I adding yet more mundane, unoriginal "content" to the cesspool?

Yikes, I think my cynicism crept through a little bit there. But yes, I do wonder whether I am creating anything that anybody would find useful. I even wonder whether the way my episodes sound, and the editing, is too generic (my episodes typically flow like this: loud intro music -> fade out to introduction of guest -> interview -> fade in to loud outro music -> tell people to subscribe). I’m still trying to answer these questions, and still trying to think of ways to make the show more interesting. I’ve thought about doing “mock job interviews”. I’ve thought about doing “podcast office hours”. If you have anything you want to hear, please let me know. There’s no point continuing to do this if there isn’t an interested audience.

Closing Thoughts

I enjoy doing the podcast. It’s fun and gets easier with each episode. The first few episodes were pretty difficult, from trying to be a good interviewer, to editing the episode down to a reasonable length. But I find that with each recording I’m learning new tricks to make my job easier, such as not being afraid to leave some dead air time while I think of something more coherent to say (because dead air is easier to edit out than my random mumbling and tangents while I try to settle on what I want to say), or setting a firm time limit so that conversations don’t meander, which makes editing more difficult. I’d like to continue doing the podcast for as long as it makes sense. I just don’t want to be unoriginal.