Expert Interview: Eric Weber Talks Data Science
As I delve deeper and deeper into the intriguing world of data science, I noticed this name, Eric Weber, frequently appearing in my LinkedIn feed. I started reading his posts and there is a true sense of kindness, genuineness, and just down right smarts.
I wanted to get to know Eric a bit more, so I asked him a few questions. He was very kind to oblige and respond. Thank you very much, Eric!
In the spirit of Eric's openness to sharing and contributing to the community, it's only appropriate that I do the same with his responses.
You're a data scientist supporting sales operations for LinkedIn Learning. Can you talk about how data science is helping this part of the business? For example, are you building applications to automate processes or are you performing more analysis/project type work?
ERIC: At LinkedIn, we take a data driven approach to sales that combines both field knowledge and data science techniques. In some cases, this entails building machine learning models to prospect for customers; in some cases, we are building scalable data pipelines to measure important characteristics of our business; and in other cases, we are doing ad hoc analysis that entails deep analysis for specific asks from sales.
In a recent LinkedIn post, you talked about the tools that you use tools, which was more along the lines of languages and editors. Are there other tools that you use for data manipulation or data visualization that you use?
ERIC: The tools I use are often project dependent. However, I’d say the tools that arise repeatedly are R, Python, SQL, Hive and Spark (data manipulation/analysis) and R and Tableau (data visualization). But otherwise, everything is built internally and I can’t go into a lot of detail about that, for obvious reasons. We are fortunate to have amazing engineering and data teams that allow us to have these platforms.
Are there aspects of the job that you find boring or tedious?
ERIC: I don’t think there are many boring parts of the job. I tend to try to automate things that I see myself repeating more than 3 or 4 times. I think if a data scientist is spending too much time on tedious or boring things, he/she is probably not being used very effectively by the organization. With that said, I find data transformation and pipelines to be fascinating so perhaps my perspective is not representative of all data scientists!
Yes, I think your perspective is different! ;)
If you see yourself repeating something and automate that; how do you share that amongst your team? Surely, they must encounter similar tasks.
ERIC: At LinkedIn, we are empowered to individually address those issues and it is part of the emphasis we place on craftsmanship. Put another way, we see it as an individual responsibility to automate repetitive tasks.
What is the most under-rated skill that you believe a Data Scientist needs to have? In other words, what is a skill that you believe many Data Scientists lack, but if they can acquire that skill, it could really benefit them?
ERIC: The most under-rated skill is writing. Writing is the vehicle by which we communicate most of what we do to others. Writing is how we shape the perception of what we do in data science to our business partners. This could be as short as an internal message, as basic as an email or as detailed as a report. In any case, writing and presentation of ideas is a fundamental skill that often is not identified as “core” to data science. My opinion is that it must be!
You’ve built quite a following on LinkedIn, did you have this following prior to starting at LinkedIn?
ERIC: I had never posted on LinkedIn until I started working here. It began as a way to share some basic updates about things I encountered at work and has taken off from there. I believe when I started I had under 500 followers. It has been surprising to see the interest and interaction I’ve generated!
What prompted you to begin sharing these updates? Did you have a desired outcome when you started?
ERIC: I started posting because I wanted to share about what I was learning at LinkedIn as a data scientist. I had no other aim, not even the aim of building a real audience.
What benefits, personally and/or professionally, has this following brought to you?
ERIC: As far as benefits go, primarily, I have been able to meet and learn from people I would never have otherwise had a chance to encounter! I’ve had discussions with people I admire a great deal and find it amazing they care about my opinion!
I 100% agree. I've been much more active on LinkedIn myself recently. It is great to draw on other sources to learn from and see things from different perspectives.
ERIC: I have had my eyes opened to issues I never knew were present and have been able to contribute to conversations I would never have otherwise seen.
What would you say to organizations that have been slow or have yet to adopt data science practices in-house?
ERIC: Two things. First, data science has a lot of potential value. That is not guaranteed value for every company. Do not rush into data science just because you see others doing it. Be principled about it! Second, it is absolutely worth the effort to evaluate the potential impact that data science can bring to your organization.
What opportunities in data science excite you in the near future and beyond? Where do you see data science heading?
ERIC: The opportunities that excite me most are in areas that deliver real value to people. I believe in a place like LinkedIn because it can transform one’s life and career. I look for opportunities that allow a data scientist to leave a lasting impact on people. I see data science continuing to grow, but in specializations. I think we are to the point where we will see sales data scientists, marketing data scientists, etc. Put another way, data scientists with specific knowledge in fields like finance, product and security will become more common. Right now we are at a very “generalist” stage. That is changing.
Thank you again, Eric. And thank you for your continued contributions to the community.