For marine ecologists, late-night work doesn’t just happen in the office. Squid come to the surface when it’s dark. To research them during my Ph.D., I worked on ships from sunset to sunrise, attracting squid with glow-in-the dark lures.
After two years in grad school and many late nights catching and releasing Humboldt squid as big as I am, I finally retrieved a pop-up satellite archival tag filled with ocean depth and temperature data that might help untangle whether climate change was the reason those creatures were migrating northward. But I was overwhelmed to realize the file was too big for Excel, the only data analysis software I knew. This was not “big data,” but it felt like big data to me because it was too big for the skill sets I had.
I’m not alone. Many scientists struggle with data, even as data has become increasingly important to science. The amount and types of data we now collect can outpace our ability to manage, analyze and collaborate with that data. One NASA project investigating how climate change affects melting sea ice relies on data from satellites combined with ships, buoys, airplanes and gliders. Many scientists haven’t been trained in the coding and workflow skills to handle this level of data—or even a single tag from a Humboldt squid.
The challenge is both building technical skills and creating a culture where we can continue to learn throughout our careers as technologies change. Every researcher knows to review prior studies before launching their own and to talk ideas through with colleagues. But when it comes to data, we don’t discuss our challenges, and academic culture often pushes us to figure out problems alone. We are rarely taught to reuse existing coding scripts on open-source platforms like GitHub so that we can build on the data science work others have already done, nor do we always have time as part of our jobs to learn to share our code there, too.
To take advantage of the amazing and ever-evolving data and tools available today, we need cultural norms that encourage us to learn from and teach one another, share data and skills, and work collaboratively to solve problems.
Open science is a movement to make research and knowledge creation more accessible and collaborative. It spans disciplines and continents. NASA staff who support researchers using NASA Earthdata are collaboratively creating open tutorials on how to use cloud computing to understand our home planet. NOAA fisheries scientists who depend on each other to flag safety concerns while sampling in rough seas are bringing that spirit of trust to their data analysis—in part through regular peer-learning meetups where they co-create open resources and ensure everyone feels psychologically safe to participate.
This growing open science movement is filling critical gaps in researchers’ data science skills. When resources and skills are shared, and asking for help is treated as a sign of intellectual strength, the result is better science and higher morale. Collaboration reduces redundancies and frustrations and helps us arrive at solutions more quickly. Scientists feel less stuck and better equipped to tackle research challenges.
In my case, patient and generous mentors helped me learn how to code and introduced me to open science communities for the programming language R, and I built confidence with technology in a way I had never imagined. I felt like Luke Skywalker learning that the Force was not only a better way to get his X-wing spacecraft out of the swamp, but that he could learn it himself, and it would then broaden his mind to what he was capable of.
Open science improved my work as a marine ecologist so profoundly that I refocused my career to empowering others and growing the movement. I founded an organization called Openscapes and built it together with fellow researcher Erin Robinson. We believe that open science is kinder science, that empathy creates innovation rather than competing with it and that investing time to build collaborations and make data and methods shareable pays off in cascading and reinforcing ways.
At a recent talk at the Women in Data Science annual conference, I shared some of the ways Openscapes supports research organizations to shift culture and promote collaborative mind-sets while also building technical skills. A first step for many research teams is to create a space and place where people can ask questions and become acquainted with data resources as a paid part of their jobs. From there, Openscapes helps organizations invest in a culture of learning that can support open science practices such as documentation.
Like the coding scripts that scientists can repurpose, we have recommendations that research groups and managers can use to incorporate open science with their own teams. For example, research groups can improve their data management practices, such as storing raw data separately from analyses. Rather than imposing top-down change, managers can identify and elevate people within their own institutions who already employ open science practices and can help foster culture change.
In the United States, the Biden-Harris administration has declared 2023 the Year of Open Science, recognizing that innovation is fueled by learning and collaboration as much as it is by new technologies and data sources. I invite you to join our growing open science movement. No step is too small—from discussing data with colleagues to exploring online communities that offer hands-on resources to signing up for a coding workshop through your university library. Together, we can learn from one another, advance truth, sustain ocean ecosystems and improve lives.