New York Times Data Analyst Interview Experience — Five Rounds, Three References, Then Rejected

It's been 4 years since this interview without my even noticing, and along with wishing everyone a happy new year, I figured I'd share the general interview process for this company.

I went through 5 rounds of interviews with this company in total, and the process was pretty drawn out. At the end they also wanted me to track down three references (a peer, a direct manager, and a manager I had collaborated with). After all the trouble of finding them, they came back and told me they'd gone with someone else. Damn it.

It started with NYT's own recruiter reaching out to me on LinkedIn. The job responsibilities themselves weren't very clearly defined. After interviewing with a senior analyst, a project manager, a senior manager, a director, and an executive director, I realized this job was full of miscellaneous tasks. The interviews covered coding, A/B testing, scenario problems, and the usual "do you have any questions for me" at the end.

For the coding portion I referenced some previous interview reports I'd found. At the time they were still using Google Cloud.

1. How do you find the unique content views for mobile (phone + tablet) versus desktop? Is there a difference between the two view counts? (I used case when + count distinct(content_id))
2. What are the peak usage hours for the two device types? You need the hour function on the timestamp. Before the interview they send you a syntax reference for the date/time functions, so read that carefully.
3. How many unique pieces of content were viewed, and how many have null values? This one is also count distinct + count(id) where is null.
4. What's the distribution of content types for the first URL each agent views each day?

Hope this information is useful for everyone.