“I saw this and thought your lab would be great for it!” - This is the text I received back in October of 2023.
It was a call for proposals from early-career researchers for a two and half day hackathon event hosted by the University of Washington and funded by the National Science Foundation - The ARTIS Exchange. Intrigued, I dug a little deeper and found out that the hackathon was designed to showcase a brand new dataset that provides global trade flows of seafood products resolved to the species level from 1996 to 2020. A dataset like this has never existed before. This is big. And it sounded fun. Unable to resist, I shared it with the rest of my team at emLab to see if anyone else was interested. As it turns out, a small group of us answered the call to action: myself, Sara Orofino - a project researcher at emLab at the time, Kaiwen Wang - an economics post-doc at the time, and Gavin McDonald - a senior project researcher at emLab.
That’s where our story begins, with a group of researchers with different backgrounds, skill sets, and research interests who came together just for the sake of nerding out over a cool new dataset. We got to work bouncing ideas off one another and eventually decided to propose a research project spinning off of some previous work Sara and I helped with. We wanted to know: do shark finning regulations (regulations that make it illegal to catch a shark, remove its fins, and throw the carcass back into the water) influence the global trade of shark products?
Not long after, Sara, Kaiwen, and I packed our bags for a trip to the University of Washington to meet the dataset creators – Jessica Gephart, Althea Marks, and Alex Godwin – and a brilliant cohort of fellow researchers.
Caption: From front, Kaiwen Wang, Echelle Burns, and Sara Orofino at The ARTIS Exchange.
The sheer diversity of projects in the room was staggering. While we were focused on shark finning, the room was buzzing with other massive questions: How do natural disasters and pandemics influence the seafood supply chain for Pacific Island Countries and Territories? How prominent are discrepancies between values reported for production and exports in small-scale fisheries? What are the patterns of trade for seafood byproducts that are often overlooked, like fishmeal? What drives the serial exploitation of luxurious animal products, like abalone? How much are nationally threatened species exported as seafood?
We immediately set to work learning the nuances of the massive dataset, eager to make strides towards answering our research questions in our short time together. We learned that this dataset combines production data from the Food and Agriculture Organization of the United Nations with trade flows from the Centre d'Etudes Prospectives et d'Informations Internationales Base pour l’Analyse du Commerce International (CEPII BACI) database. It provides annual estimates for exports, imports, and consumption by country and commodity-code using a mass balance approach to bridge production and trade data. Through reconciling production and trade data, this dataset estimates the live weight equivalents of traded primary products. In other words, it answers the question: how many sharks were actually needed to export this many tonnes of shark meat fillets?
As we zeroed in on our project, the data forced us to get strategic. Our first decision was around what commodity-code version to use since ARTIS includes every version from 1950-2017. We needed enough data before and after regulations were implemented. Because most shark finning regulations were implemented in 2012, we decided to use the 2007 version of the commodity codes and narrow our scope to 2007-2020. We also had to pivot around a massive historical data gap: “shark fins” and “shark meat” weren’t separated into their own product codes until 2017, meaning we’d have to track the footprint of finning regulations through the broader category of shark meat, which included shark fins. We were lucky to have the database creators in the room while we were applying the database directly to our research question. Even still, we were left with more questions than answers by the end of the hackathon, but we packed our bags excited to continue this work together. Back home, the real challenge began: trying to push forward with our passion project while still hitting the deadlines for the many other projects that demanded our time and attention.
Caption: Shark product trade flows (mt) before (2007), during (2014), and after (2020) major regulation adoption. Arrows show trade direction; bold lines highlight top 10 annual flows. Country shading reflects regulation strength (Protected - sharks are fully protected from fishing; FNA - fins naturally attached; FCR - fin to carcass ratio; FRU - unspecified finning regulations). The world basemap was pulled from the ‘rnaturalearth‘ R package.
Over the coming months – and eventually years – we traded off on pushing the project forward. Sara and I worked on data cleaning, visualization, and code organization, Kaiwen was our resident economist who spearheaded the primary analyses, and Gavin was our fearless overseer. We also brought in collaborators and shark experts that we had worked with in previous projects to provide their perspectives on our work. We continued to make progress, even as Sara moved on to another position at The Nature Conservancy and Kaiwen to a new position at Renmin University of China, but eventually we found ourselves at a standstill.
We had used a difference-in-difference approach to compare trade outcomes – domestic exports, foreign imports, domestic consumption, number of trading partners. We took the difference in trade outcomes before and after the implementation of shark finning regulations among countries with regulations, then compared that difference to the trends of countries without regulations. We expected the data to tell a clear story. Maybe shark finning regulations deterred fishers from catching sharks, causing exports to plummet. Or maybe fishers pivoted to selling shark meat alongside the fins, causing exports to spike.
But we saw… nothing.
There was no statistically significant effect. Kaiwen ran a slew of robustness checks to ensure our model and data were sound, but the null result remained. The only faint glimmer was among a few early adopter countries: after seven or more years of implementation, they began exporting less and importing more, effectively becoming less dominant as exporters in the global trade of shark products. Still, we just didn’t have enough data for these results to show true statistical significance; we needed a longer time series.
While disappointing, we knew that this work was still important. This was a first-of-its-kind approximation of how fisher behavior impacts global trade using the ARTIS dataset, and our methodologies could easily be applied to other sectors, other regulations, and other species. So, we pushed forward. We brainstormed alternative hypotheses, looked more closely at the data, and eventually channeled our findings into a manuscript. After a relentless cycle of submissions and revisions (you know how it goes), our perseverance paid off. Two years after the hackathon, Nature Communications recognized the value of this first step towards a better understanding of regulations and trade outcomes and accepted our paper.
If you had told us during that sprint in Seattle that our hackathon project would eventually end up in Nature Communications, we might have laughed. But after two years of grit, countless robustness checks, and a relentless revision process, we are thrilled to present our peer-reviewed journal article: “Global trade responses to shark finning regulations”.
We’re also not the only group from the hackathon who made it to the publishing world - Fisheries Trade and Blue Nutrient Flows in Pacific Island Countries was published in Fish and Fisheries in 2026, The growing role of trimmings and their origins in global fishmeal production and trade was published in Environmental Research in 2026, and Trade and governance drive luxury seafood serial exploitation is under review in Nature Portfolio. It’s exciting to see how much work has come out of this one event, and we’re only just scratching the surface on how this dataset can be used.
