Before joining the City of Boston, I worked at Oceana, a global ocean conservation NGO, using data science and analytic techniques from particle physics to tackle the challenge of illegal fishing:

​

​

We are facing an ocean health crisis.  Climate change, ocean acidification, pollution, and overfishing are stretching marine ecosystems to the breaking point, threatening their ability to sustainably provide food and livelihood to billions of people around the world.  Rampant overfishing has imperiled global commercial fish stock to the point that nearly 90% of the world’s fisheries are already fully exploited, over-exploited, or depleted to the point of collapse.  Illegal, unreported, and unregulated (IUU) fishing is a major contributor to this crisis, accounting for an estimated 20% of the total catch and costing tens of billions of dollars annually.  The perpetrators of IUU fishing ignore protected areas, use harmful and invasive fishing techniques, follow no standards of health and sanitation, and are often associated with other illegal activities such as forced labor and even human trafficking.

​

By nature, combating IUU fishing is a formidable challenge.  It often takes place outside of governmental jurisdiction on the high seas or in the waters of developing countries that lack the resources to enforce the law.  Illegal fishing vessels change names, owners, and flags frequently, and will turn off or manipulate required tracking systems to obscure their locations.  The fish is transported by third-party reefer ships and brought to market hidden under a false species designation.  Licensing and other documentation is forged and money earned disappears behind the borders of haven countries.  Even efforts to define the scope of the problem have been confounded by this complex tangle of issues and its globalized nature.

​

Particle physics and IUU fishing might seem like an unorthodox partnership, but the two fields share ample common ground.  The quantity of data produced in particle collisions is staggering: the data center at CERN stores more than 30 petabytes of data per year, a figure comparable to 250 years of high-definition video.  As a result, particle physicists have developed sophisticated techniques to sift through large and unwieldy datasets, manage and catalog the important pieces, and uncover patterns beneath the noise.  These techniques can be usefully applied to data outside the realm of particle collisions.  In addition, collisions often produce particles, such as neutrinos, that cannot be detected by the experiment’s detectors.  Instead their properties and behavior must be inferred by the properties and behavior of other output particles.  Some particles are so rare that they only appear a handful of times in billions of collisions, so analysis of their properties and behavior relies on extracting a story from very limited information.  Physicists have developed many techniques for addressing the problem of working with partial or limited data, techniques that transfer well to IUU fishing where all data has inherent caveats and limitations.