#124 The Data Science Team: Skills Needed, Purpose and How to Structure with Dan Costanza – Chief Data Scientist
Dan Costanza, Chief Data Scientist at Citi, joins me for the first episode in the new “Bitesize Insights for Data Driven Leaders” Series. Dan opens the show by explaining how he got interested in data. After graduating from college, Dan went into an investment banking role. Eventually, he received an exciting project that got him started down the data path. Dan says as someone who didn’t study computer science in school, it has been a heavy lift trying to get those technical skills up to par. During a code-heavy project, Dan needed to learn how to break up the project and work through it. Also, he learned how to think about sampling data without bias.
Then, Dan explains the importance of emotional intelligence for data science. Conscientiousness and emotional intelligence are the things that you can actually interview for. Instead of judging people on their grades, we need to judge people on their ethics, communication skills, and willingness to work in teams. In India, Dan set up a data science team. The talent in India is insane. However, there are cultural differences Dan needed to work through. For instance, he told his team that they needed to speak up when they had ideas. If you create space for people to bring their own thoughts, you’ll hear loads of good suggestions. Before Dan told his team that, they would withhold useful information.
Later, Dan explains when a company can determine the need for a centralized team. First, it depends on the type of work that you do to determine if a centralized team is necessary. There are two types of data work, predictive and descriptive. The more predictive you are, the more expert work you need. If you are trying to predict whether or not a stock will go up, then that data science work will be challenging. However, if you are trying to solve day to day business problems, standard models can be applied. Plus, Dan says you need to have a centralized team at the beginning because you need someone to set up servers. Stay tuned as Dan gives his opinion on data governance, the importance of leading by example, and Dan describes his cadence on projects during the week.
Enjoy the show!
We speak about:
[02:30] How did you get interested in data?
[06:50] Why did you decide to go into investment banking?
[08:45] How does emotional intelligence play a role in data science?
[11:15] How was setting up a team in India?
[17:00] Is there a point when a centralized team is better? Is there a point when a transition is necessary?
[20:30] How do you pick the earlier divisions?
[21:45] How does predictive vs. descriptive work within the team?
[23:20] How do you see the use of data governance?
[25:45] How do you ensure people are looking for bias and fairness in their data?
[26:50] What is your cadence on projects during the week?
Resources:
Dan’s LinkedIn: https://www.linkedin.com/in/daniel-costanza-73882814/
Citi’s LinkedIn: https://www.linkedin.com/company/citi/
Quotes:
“When you look at the hiring research again, like there are two real categories and the one is things you don’t interview for, which are the intellectual horsepower things and those are - how smart you are, do you have some specific skills I need. The word that always comes up on the other side is conscientiousness, and that encompasses the stuff we talked about at the beginning, and the emotional intelligence, teamwork parts of it and those are the things you do actually interview for. Which is counterintuitive for a lot of people who work in quanti type roles because you want to ask people really hard questions, to see if they are smart, but the problem is the data doesn’t support that as being predictive of anything when you control for their grades.”
“You start by spraying things around, working with a lot of people, just to get the volume in and see who those people are and meet people, and as you work a little bit, you start to understand their own types of workflow.”
“More powerful then compliant is having good ethics there on the ground.”
“It was a pretty incredible experience because the talent we were able to recruit there was just truly insane. The example I give - we used to give people this placement test there to prescreen who we would interview, and it was a 30 question math test we would give to each candidate, and I took it on a whim just to see how I would do and got a straight zero. Did not answer a single question!”
We are now on YouTube! Watch the episode here: https://youtu.be/IHaYicKq_8g
And as always, we appreciate your Reviews, Follows, Likes, Shares and Ratings. It really helps new data scientists find us.
Thank you so much, and enjoy the show!