Unfortunately I was impacted by a layoff at my former company, Snapdocs, earlier this year. After taking a short break, I started a full throttle search for a new role. The job hunt definitely is difficult in this current atmoshpere. Given all the other layoffs and the dwindling openings available, I knew it was going to take time and determination. It was a full time job itself, working on networking, searching for jobs, applying for jobs (with customized resumes and cover letters), and preparing for interviews. The first couple interview cycles were a little rough with some rustiness on my part. But I started regaining that interviewing muscle memory and started not only getting more interview requests but getting further into the interview rounds.

Going into the job search, I was hoping to accept an exciting opportunity within six months. I knew that I my goal was not to just accept the first offer I got. I have some things I look for and certain non-negotiables. Consequently, I was unafraid to ultimately pass on a job offer if it did not look like it would be a great fit, for both sides. Its important to me that there is that good fit for both myself and future emmployer to ensure success for everyone. But I found an awesome role with Sprout Social! I consider myself very lucky to be able to get such an exciting role, very much breaking my initial six month expectation. In fact, from layoff to accepting the offer, it was only 84 days! I am very grateful to the numerous people who helped my directly and indirectly in this search.

Details and Viz

Of course, like many other things in my life, I kept track of the job search and thus have a few visuals to share for this post. I have seen Sankey diagrams to show the job search outcomes and wanted to make one as well, which you can see below. A few things stick out, at least to me:

• The overall response rate to my 148 job applications was a little over 12%. At the beginning of the search, this percentage was much lower and steadily increased as time went on.
• Nearly half (45%) of my applications were rejected with an automated email and a slightly smaller percentage (42%) never responded at all.
• Of the 18 applications that I started an interview cycle, 5 ghosted me and I still have not heard from them. Still other companies also did not respond for weeks after the final round before sending an automated rejection email.

In addition to keeping track of application statuses and the like, I also kept track of all the actual job descriptions. For these, I wanted to create a word cloud (see it below). What are the common keywords across all the data science and machine learning roles that I had applied to? To do this, I applied a homebrew version of TF-IDF to get the word counts before visualizing. I removed traditional stopwords and added several to that list after looking at the first output of the algorithm. The big factor in the additional stopwords were numbers (i.e. “5 years of experience”). After a few iterations, the final output is below:

Its not as obvious what the top ten words are, but for easier reading, those are:

1. data
2. work
3. team
4. product
5. python
6. model
7. ability
8. business
9. machine learning
10. data science

It should be obvious that “data” would be the first and most popular term. What is more interesting to me is that “product” ranked higher than “python”, “model”, “data science”, and “machine learning”. In addition the term “business” also is in the top 8 beating out “machine learning” and “data science”. This should not be too much of a surprise as a (non research) data scientist should be focused on business goals and objectives. After all, the point in a company is usually to solve business problems and not necessarily academic ones.

Parting Thoughts

I wanted to make a quick post about my job search from this year. I collected and tracked a fair amount of data and it was fun to visualize it to better understand that journey. Please do not hesitate to reach out and contact me if you have any questions or if I can help in any way. Cheers!