How a walk on the beach taught me to be a better data science manager

How a walk on the beach taught me to be a better data science manager

Like sand beneath the sea, data seems endless. If you’ve ever had the privilege to walk alone on a beach perhaps you’ve gone through these motions. You take in the tempest of the sea or lack thereof. The peaceful sound of the ebb and tide or crashing waves. A seagull or sandpiper picking at muscle shells. The raging wind found in Dingle Bay or the sweet plumeria scented breeze off Oahu’s coast. Occasionally you can find unique rocks or shells. There’s one shell that scientists are particularly fond of: the conch shell. The intricate spiral moving closer and closer to the center point.

blue sea shell illustration

Keep that picture of the conch shell in mind while we open up our laptops, where endless streams of data waiting behind each of our dashboards. Decisions are being made and deals are being done based upon the information streaming onto your screen. It can seem endless and chaotic if you don’t see the pattern of movement.

As a woman in STEAM, data science and analytics was a natural part of my vocation. Even though I have been in the Seattle Tech Hub for the past seven years, it wasn’t until I started working at Neal Analytics that I thought about how data is displayed in front of me to make sense.

It isn’t magic – even though it often seems like it. Previously, I had been on the receiving side of data analytics. Now, I was on the producing side. The teams I work with at Neal Analytics are the real magicians who build data models, pipelines, and dashboards. As a technical project manager, it wasn’t until recently that I was introduced to the infinite possibilities of data analytics. With my team’s help, I have learned how to better explain the timing around data science projects for stakeholders, executives, and the wider business audience. Their process and insight teach me how to focus our questions and fit the projects to our clients’ needs.

Why do data science projects (seem to) take so long?

I have witnessed our own data scientists spend months working on model hypotheses and theories. They dance with equations, moving back and forth and all around in order to predict customer trends and examine how external circumstances such as health, fashion, and weather can impact revenue.

The data science wheels can seem to spin endlessly with repetitive testing for months with ever more predictive analyses opportunities and model possibilities. Businesses need road maps with realistic timelines and attainable goals in order to work towards profitable quarters; how can potentially months of data science prove to be profitable?

Mapping progress in business and data science

How do we plan data science efforts and quantify it in ways business and marketing agents can understand?

It all comes back to the walk on the beach. Billions of strands of data, like sand, and the analytics and questions we propose are circling round and round like the tunnels of the conch shell, with its Fibonacci* sequence curving us ever closer to a conclusion.

sea shell diagram for data science process

A similar type of movement is found in data analytics. We build models and run tests from a single point, with a specific set of data. Then we move to another point around the curve, so to speak, running the tests again and reviewing results, and adjusting accordingly. From each vantage point, we see a unique perspective. We string this information together and it allows us to make more informed decisions.

I have now come to understand the bigger picture of data science as a multi-faceted project. Data scientists must look at information from multiple angles to fine-tune for accuracy while incorporating the client’s chosen factors.

Take this real-life scenario for example I was managing a team that needed to analyze a client’s sales revenue over several years. They would analyze one week’s worth of customers in and out of the store and ask questions to narrow down the expectations. What did we need to drill down on? What was the larger objective?

From there, they would move to analyze a month of sales revenue data. Then a quarter. Then a year. Each stage allowed for clearer information and broader analyses until we finally reached the “epicenter” where results, insights, and knowledge met to prove or disprove the hypothesis.

To me, that epicenter is like the center of the conch shell.

The natural mathematical design in the conch shell continues to help me help my clients develop successful roadmaps with data science. Once progress can be seen from a different angle, a spiraling drill instead of the usual linear timeline, each side can get the best work from the other.

Perhaps the next time you are pondering the value of data science, the endless possibilities of information, and how you can use it to provide value for your company, you ought to take a walk on the beach.

* Our in-house data scientists would like to make note that this analogy with the conch shell’s Fibonacci sequence is not a direct correlation to the data science process. They did, however, agree that it could still be used to abstractly demonstrate the overall point.

For more information about data science: