Skip to main content Skip to secondary navigation

Spotlight: The devil is in the data details

Main content start


Shinnosuke Nakayama

Data Research Scientist

Published January 29, 2024

I was always interested in collective behavior. Fisheries just happened to be my choice of research topic.

My hometown in Japan is not by the ocean, but my mother’s is. We stayed there every summer and I swam at the beach. We would visit the aquarium in Osaka, which had a whale shark swimming elegantly in the tank, and sardines swirling in the water. I remember being absolutely mesmerized by the collective motions. I was about 12 years old at the time.

I studied fish behavior as an undergraduate and for my PhD, and then moved on to study engineering and human behavior, specifically human-computer interactions. My main focus was to understand how humans interact with each other through computers and change their behavior in the presence of peers. I joined the Center for Ocean Solutions because it combined all of my academic pursuits: data science, oceans, and human behavior.

Data scientists deal with the whole life cycle of data, from designing data collection to processing, analyzing, and delivering data. Many people think that as a data scientist, I’m doing state-of-the-art analysis or visualization, but 99% of my time is spent cleaning up data. 

I always like being involved at early stages of discussion, when researchers determine the kinds of questions they want to ask – even before they know what data they want to use or collect. If you set the team up to collect “clean” data, it’s very easy to analyze later. 

In my role at the Center for Ocean Solutions, I bring a data science perspective to other researchers focused on policy and the social sciences. For example, consider illegal, unregulated, and unreported fishing, or IUU fishing. While there are numerous ways to address illegal fishing, exploring all possibilities can be inefficient or impossible. Data science helps the team identify the most promising approaches for ending IUU fishing.

For example, through machine learning, we can analyze large volumes of satellite data to piece together patterns of vessel movements and behavior. This analysis allows us to identify regions and ports at highest risk of illegal fishing, and identify specific risk factors like the country a vessel is registered to and the type of fishing gear it carries onboard. These data-backed insights help us hone in on possible solutions to illegal fishing.

What I like about the Center for Ocean Solutions is that we work with NGOs, businesses, and governments. Sometimes our results are presented in international meetings. I like that because it means you’re going beyond publishing a paper and having only like-minded academics read your work. I believe you need data in order to say something meaningful. People need proof points. 

I try to digest the data as much as possible so non-data science people can also understand the message. For example, people who can influence marine policy may have difficulty understanding complex data. We need to make more room for data science in forums where decisions are made about shared natural resources like the ocean.