By: Yadira Y. Caro
Data Science has become very popular term in the world of technology careers. But what does this term really mean? How can you start shifting your skills to become a data scientist? Kristen Kehrer wants to help with that.
With a Bachelors in Mathematics and a Masters in Statistics, Kristen has worked in fields such as Health, Communications and eCommerce. Her roles have included analyzing data, conducting research and developing technical models as coder. When she started, she did not knew these were roles would be ascribed to a Data Scientist.
Today, as a founder of Data Moves Me, she focuses on teaching others about the field through online courses, speaking engagements and helping people build their resume towards a job in Data Science. She is also a Data Science instructor at UC Berkley Extension and EMERITUS Institute of Management.
In this interview she describes what Data Science is and shares some of the required skills to get into this career.
How do you describe Data Science and what you do?
This completely depends on the context and who I’m talking to. The definition I typically use for data science is: “It is the understanding and utilization of tools, data and methodologies that enable you to effectively solve problems utilizing data.” Someone who self identifies as a “data scientist” is often using machine learning and writing code, however the umbrella of the “data sciences” also involves analysts and other data wranglers.
It is certainly a multi-disciplinary field including a bit from programming, statistics, and business. There are no unicorns, everyone has their own strengths in the field and may be doing quite different tasks depending on industry.
What are some of most common misconceptions about it?
Again, the misconceptions depend on who you’re talking to. There are people who think everything is “AI”, there are the people who aren’t as data literate but still making decisions based on data, potentially the most dangerous (people). There are those who do not understand what the real pipeline looks like and only focus on machine learning.
I think there are a whole lot of misconceptions and it’s exacerbated by the “hotness” of the field. Lots of buzzwords and hype that make it difficult for people to fully grasp what the reality looks like. There is a huge focus on machine learning, but this is one tool.
I often hear people say “I need to hire a data scientist.” This is an incredibly broad statement. Think first about what you really need someone to help you with, nail down how they’ll contribute to strategy and what skills that will actually require, and then hire for those specific things, rather than listing the kitchen sink in terms of skills on a job description.
Why did you choose it as a career?
I definitely didn’t know that I was seeking out “data science.” The term wasn’t really being used when I started my career. I had finished a BS in Mathematics in 2004, realized I was in a dead-end job and decided to go back and pursue a MS in Statistics. I had seen that statisticians made good money. Then it was through a series of job changes and career moves that I really found myself in the data science space. It also involved some rebranding, as I considered myself a statistician who does “advanced analytics”. Then one day it was “oh wait, I’m a data scientist”.
Can you describe a project you worked on which you enjoyed or learned from?
The amazing thing about this field is that I’ve found most of the projects enjoyable. This industry requires continuous learning. Even after I’ve implemented an algorithm one way, the next time I go to do something similar there is probably a new library or package that makes data cleaning or model building easier, so I learn those.
One of my more favorite projects was using customer’s subscription data to find customers with seasonal usage patterns. So instead of saying “hey, these customers are using our product less and may be a retention risk,” I was able to say “hey, this customer has a seasonal business and we expect less usage from them in these months, we can use this information to speak to them differently and infer there needs.”
I used the TBATS algorithm to take these people as seasonal or non-seasonal. Although I’m very well versed in econometric time series analysis and forecasting, this was my first time researching this algorithm and the pros and cons that went along with it. It was also sort of an off-label use case for the algorithm. That is where I find the most enjoyment: developing a methodology that will work for a problem I haven’t solved before.
Because Data Science is so interdisciplinary, there are many competencies that transfer well from other careers if you position them for the Data Scientist role. I want to educate others to be able to use this to their advantage.Kristen Kehrer
What drove you to focus on helping others with resume building?
I was laid-off in 2017 a week and half after returning from my second maternity leave. Although I was quite happy with my resume as is and was frequently getting calls from recruiters, I picked up some amazing additional tips from a career coach. I saw so many people trying to “rebrand” themselves or make a career change to data science. These people would ask me to review their resume and it was clear that they were highlighting the things they had done previously, but not how that would translate to them being an effective Data Scientist.
Because Data Science is so interdisciplinary, there are many competencies that transfer well from other careers if you position them for the Data Scientist role. I want to educate others to be able to use this to their advantage. People often bring fantastic skills to the table that they’re not highlighting to their full potential.
What trends do you see coming up in the field?
Well I hope that there will be more of a standardization between terminology, roles and responsibilities so that we can all use a common language and understand each other. I think as Data Science matures it will be clearer that it is a team sport and not a single person sport.
What are two of the absolute must-have tools you use in your day-to-day for your job?
I always say that SQL is a must. People often get distracted by shiny objects, learning new algorithms, etc. But on your first day as a data scientist you’ll most likely be told about your new job’s data warehouse. That is where you’ll extract your data from. Although you can do joins and connect to a database from R or Python, you’ll still need to understand relational databases to navigate the schema where your data lives to be effective.
I also like to stress the importance of communication skills. Give your deliverables love and care, think about how to best present to a non-technical audience. Your ability to build relationships where stakeholders trust your work and see you as a valued partner will be instrumental in your career.
What are some of the plans for near future?
I’m currently working on a book Mothers of Data Science with Kate Strachnyi. I expect the book to be available in 2020. I’m also teaching a course through UC Berkeley Extension called “Practical Data Science.” This is a foundations of Data Science course in R. I’m also currently consulting and offering in-office training for Analytics/Data Science teams that want to take their skills to the next level. I also intend to keep blogging at https://datamovesme.com.
What resources (books, podcast, websites, etc.), do you recommend which have helped you in your career?
I try to give useful tips on my personal blog https://datamovesme.com. I also think finding a community on LinkedIn, Twitter, or other social platform helps you to keep up with the trends, new programming libraries that will make your life a little easier, and help you to gauge what might be most relevant to learn next. Because again, it is continuous lifelong learning as a Data Scientist that will help you stay relevant. You can also become involved in things like “Makeover Monday” or “Tidy Tuesday” and the community will give you feedback on your work. This is one of the greatest forms of visibility, and networking is diving right in and contributing.
Do you have questions, feedback or suggestions of people to interview? Contact me!