In our previous article, “Modernizing Data Science in Today’s Business” published October 30, 2018, we explored the ways in which data science has become essential for thriving modern businesses. While that article focused on the steps a company should take to become more data-oriented and the benefits it can expect to enjoy as a result, we focus this month on the flip side of the equation - what do individual people need to do in order to be successful as a data scientist? We interviewed a few of our talented data scientists: Shree Reddy, Jason Livingston and Chris Fox, about their accomplishments, experience, challenges, and insights.
Life as a Data Scientist at Fulcrum
December 11, 2018
In our previous article,“Modernizing Data Science in Today’s Business” published October 30, 2018, we explored the ways in which data science has become essential for thriving modern businesses. While that article focused on the steps a company should take to become more data-oriented and the benefits it can expect to enjoy as a result, we focus this month on the flip side of the equation - what do individual people need to do in order to be successful as a data scientist? We interviewed a few of our talented data scientists: Shree Reddy, Jason Livingston and Chris Fox, about their accomplishments, experience, challenges, and insights.
Versatile data scientists tackling the full spectrum of “data” projects
We spoke with three of Fulcrum’s data scientists whose experience includes a mix of data-driven projects ranging from statistical modeling to business-oriented data consulting. This variety reflects the trend of the widely varying application of data science in modern businesses, some examples of which are described from their point of view when we asked,“What is your flagship project or a project that was most interesting to you?”
As noted by Shree, it is not uncommon for a data scientist to have the responsibility of building a statistical model in the context of risk and/or pricing analytics, which she considers her flagship project.“The goal for us was to develop a predictive model to forecast the loss ratio for new business clients by using easily interpretable variables”.In addition to developing an interpretable model, it is crucial that the model is socialized with the key stakeholders effectively. She adds,"It is important to visualize your results so that executives understand results, observations and/or validations easily... this project has helped me understand and think on how every step you take will have an impact".
Improving a Media Mix Model (MMM) was noted as a main pillar of Jason’s flagship project. He says,“Coming from a more marketing intensive background, I had experience being on the other side and seeing how cost efficiency plays out in data science”. He adds,“My goal became to discover data that could serve as indicators of efficiency and incorporate these features to improve spend optimization”. Jason expressed excitement to be on a high-profile analytics project that can make a global impact."Improved forecasting accuracy of the media mix models provided clients with a more useful tool for projecting future sales and revenue, which are important for their public investors", says Jason.
With his years of experience at Fulcrum, Chris has participated in a wide range of data science projects. Most recently, he has been working with a data platform start-up company to support end-clients’ successful use of the platform. He notes,“When I was brought on to work with them, their company was growing rapidly and found itself backlogged with aggressively deadlined projects... the Fulcrum team worked hard to figure out and lead the way on platform usage for early clients”.He adds,“It took a lot of work to prioritize and execute, but I can say that I am very satisfied with the role we played in shaping the platform in its earliest client engagements”.
These are just a few examples to illustrate how Fulcrum’s data scientists are seeking the best way possible to help clients reach optimal solutions to maximize the value of data.
The unspoken pain points
Fulcrum’s data scientists also shared some commonly experienced data science pain points. Three common challenges can be summarized as:
- Modeling restrictions;
- The “pressure” on the data scientists and;
- Liaising between business and technical individuals.
Though model building is an essential component of applied analytics, data scientists are often faced with restrictions. More often than not, the required data isn’t adequately captured to be able to build an intended model. Shree notes from her recent predictive modeling project,“When the historical performance variables are limited, we can only refer to exogenous variables that are less reliable… hence, dealing with all data limitations was the biggest challenge”. In some cases, such data is available with third party vendors, but for a fee. Additionally, there are moments when a data scientist isn’t allowed to use certain types of information to draw conclusions as per regulatory requirements.
When an organization heavily depends on data science for their decision making, the data science personnel faces tightened turnaround time and heightened expectations. The pressure becomes even greater when data science is fully integrated in the company’s day-to-day operations.“The client urgently needed the optimal dollar allocation to fix operational issues that were making their end customers unsatisfied… insights generated from data science were not only to serve business stakeholders, but also to improve customer satisfaction”, says Jason of his MMM project. Inevitably, there was pressure to improve forecasting accuracy of the media mix model. Even for the best data scientists, such situations can be stressful, but they can also be very rewarding and create an opportunity for positive challenge.
Though it is perceived as a deeply technical field, data science is an interdisciplinary practice that requires effective communication and creative problem solving skills. For example, within Chris’ flagship data platform project, it was especially important to recognize that initial client satisfaction was critical in growing the business. He recalls,“When we arrived, this company was at risk of a number of their early clients being disappointed with their experience. We saw it as our responsibility to prevent this from happening”. He adds,“I am confident we played a crucial role in preserving some of these key relationships by establishing clarity on the core product and streamlining future sales efforts, which allowed them to establish proper client expectations upfront”.
Why Fulcrum is the best team for the job
In speaking with a few of our data scientists, each cited that collaborative teamwork, a culture of innovation, and the intersection of technical and business expertise are the components that allow Fulcrum to overcome challenges and achieve desired outcomes. In Shree’s opinion,“The best part is the amazing co-workers”. She adds,“They are simply helpful and encouraging”. Additionally, Jason notes that the culture of continuous learning and thirst for innovation are the reasons he feels Fulcrum is a go-to for hard to solve projects.“My experience at Fulcrum has been rewarding primarily in terms of learning and growth... I’ve applied modeling methods here that I never considered applying in my previous work, and as such, I’ve been able to sharpen those skills and learn several new modeling methods”. Chris cited high impact, cross-industry solutions as unique and effective aspects of working here.“Fulcrum is a great place to work as a data scientist, particularly because it gives you the chance to see how data science is used across a variety of businesses”.He adds,“Also, I was able to sharpen my non-technical skills from sitting in dozens of sales meetings and understand what clients were really looking to do with their data platforms”.
Every project Fulcrum takes on is an opportunity to showcase our talented employees and become the best at what we do. Our Data Science Acceleration Team (DSAT) is customized for each client to include the right mix of engineers, developers and data scientists like Jason, Shree, and Chris. We encourage you to follow us on Linkedin and see what we’ll be up to in the new year.