“Wait, I’m confused…are they spicy hot or, like, microwaved hot?” someone* asks.
It’s that time of our daily morning Research and Development (R&D) stand-up meeting when Rick, our COO and CFO, asks if we have anything else we’d like to discuss before we break. For whatever reason, the conversation has devolved to waxing poetic about Flamin’ Hot Cheetos, so it’s clear that this meeting should end. I make a mental note to pick up a bag on the way home tonight, possibly to share with the rest of the team at our weekly huddle.
The R&D team at Comlinkdata is a relatively small group of seven people (of which I am the most recent addition, starting my role as a Senior Analyst on the team just a few days prior to last Thanksgiving), and we function in a capacity somewhere between Data Operations (Data Ops)—who clean and validate existing and incoming data—and Client Analytics and Insight (CAI)—who present analysis and telecom industry strategies to current and potential clients. Our team is managed by Matthew, Comlinkdata’s Director of Product Development, and the analytical work we do as a team is in service to that title: we help to create the next generation of products for the company as well as improve the processes in place for existing products and methodologies. It’s a great team to be a part of if you like to balance your detail-oriented, analytical nature with a large shot of creativity! Throughout the development process, we are encouraged to “celebrate failure” in an effort to home in on our final, production-level deliverables (at which point, failures are not celebrated). This embracing of “failure” in trying out new ideas and techniques is especially evident during our weekly Wednesday huddle session, where team members present projects at various stages of completion in order to get feedback about improvements that could be made or logical next steps to take. We often invite CAI members to join the huddle so that we can get perspective on what might be most useful to our clients, too.
What does a typical day look like for me? Quite fortunately, no day so far has really been typical (I know, I know, everyone says that, but it feels true on our team). During the summer months, my day starts with a strong iced coffee (our downtown Boston office is within walking distance of lots of good options) and usually a breakfast sandwich or bagel (again, there are several nearby options). I tend to take a few minutes before stand-up to prioritize my ongoing work streams while I wait for the caffeine to kick in, but what gets top billing on any given day depends on what week of the month we are currently in.
Broadly speaking, most of the work that I do on the team falls into the following categories: analyzing and understanding consumer behaviors to help improve customer retention and acquisition, visualizing data, and addressing miscellaneous statistical/methodological inquiries. I work most closely with Tyler, a manager on the R&D team, to carry out the majority of this work, as we both have academic training in statistics. The work we put in to creating customer segment analytics for clients allows them to make better business decisions, such as how to best allot a limited advertising budget, as they can focus their advertising efforts (in whatever form that takes) on segments with high potential. Efficiency is crucial for using big data tools (SQL, R, Spark, Hive, Tableau, etc.) to complete all of this work, so much of my and the rest of the team’s time is spent streamlining our existing processes, both to expedite our daily scripts and to consider how to best scale things in the future to handle the demands of larger quantities of data. For example, several of us on the team had been generating a specific kind of chart in Excel, by hand, which was time-consuming (and possibly error-prone); I thought it might be beneficial for everyone if I simply wrote a function in R that did most of this work automatically, and it is now one of our go-to methods for model comparison. I am also a staunch believer in data visualization as a tool for exploring patterns inherent in large data sets, in addition to presenting effective final results to clients. In fact, it was through some basic graphing that we realized that one of our data sets, which is supposed to be nationally representative, contained zero entries from a particular state**, which is now something that we know needs to be corrected. Finally, I am occasionally asked to help brainstorm solutions to questions like, “What’s the best way to create a control group for this test when there isn’t or can’t be one?” All in all, life on the R&D team is challenging, dynamic, collaborative, supportive, and fun (our Slack channel post-4:30 p.m. on Fridays is a testament to this). I am excited for what the future brings for the team and for Comlinkdata!
* Name forgotten to protect the ignorant.
** Of Twilight fame.