Open Source and Open Science in Latin America
A conversation on growing global communities with local roots
Authors: Laura Ación, Gonzalo Peña-Castellanos, Fernando Pérez
At the 2021 Chan Zuckerberg Initiative (CZI) Essential Open Source Software for Science (EOSS) meeting, CZI hosted a panel with Laura Ación and Gonzalo Peña Castellanos, moderated by Fernando Pérez. This post represents reflections on the topics covered and those that weren’t due to time constraints. A longer version of this material is available at the MetaDocencia blog that includes translations to Spanish and Portuguese.
Like other areas in science and technology, scientific open source development tends to be dominated by activity originating in the U.S., Canada, and Western Europe. But for those of us who hail from Latin America, its real promise lies in offering not only access but also agency as first-class participants and co-creators to people from all nations.
This is a rich and complex topic, one that deserves both discussion and study. We are only three practitioners, so we present these ideas as a starting point for a conversation: our perspective is limited to the slices of our own countries we know well (Latin America is a large continent with vast ethnic and cultural diversity), and we don’t pretend to generalize. We hope it serves to open the discussion around how we can better make the development of open source and open science a more global endeavor.
Fernando Pérez: Can you speak to the importance of language localization as a tool to broaden the reach of open science in Latin America?
Laura Ación: Until we all can afford the time and money to learn English, I see no other way to make science inclusive. I know that considering only English as the scientific lingua franca makes humanity miss out on massive amounts of knowledge.
Gonzalo Peña Castellanos: Also, just because people in our countries are not fluent in English, there is no reason to lag behind in technology, programming, and data literacy. The argument that a scientist must know English is also exclusionary to anyone not a scientist, and we want to increase science literacy and advocacy.
Fernando: What other barriers exist for the development of open source and open science communities of practice in Latin America?
Laura: When you live in Latin America, the problems are obvious, but they are difficult to grasp when you are not within the region. Thus, it is critical to have local people with a history of advocating for the collective good at international decision-making tables. They will have a better understanding of what ideas click locally and which don’t. The more there is a dialogue between all geographies, the faster all barriers will fall.
Gonzalo: There are barriers associated with basic infrastructure, such as lack of access to stable, reliable internet with good speed and 24/7 access. Additional barriers are more systemic and relate to both human and physical infrastructure of the education system.
Fernando: You both organized R and Python events in Latin America. Is Latin America’s relationship with the rest of the open source world where you’d like it to be? If not, what could be done to create stronger global networks?
Laura: In the R ecosystem, what is still missing is a good representation of Latin America and other marginalized geographies in the decision-making positions that drive the international community. I have not seen companies, except for some isolated individuals working for companies, moving toward a meaningful inclusion of Latin America. Companies have the most prominent wallets and could make a considerable difference by sponsoring regional efforts.
Gonzalo: Things have been steadily improving, but the problem with all the volunteer work associated with regional R and Python events is that it puts a lot of strain on the organizers. Having paid positions for these advocacy roles could be an option and larger companies in the region have been investing more and more in these roles.
Fernando: Are there public policies to promote open source and open science in your countries or in other parts of Latin America? What are the greatest difficulties to carry them forward?
Laura: There are a lack of incentives to promote open source work and science in academia in our countries. Latin American researchers who promote open data, code, and science often voice their concerns about global asymmetry in open science. For example, it can take years to collect valuable data frugally, with vast amounts of hard work. The moment the dataset is open, a lab in a high-income region can get its 10 highly skilled persons to analyze the data and write the papers from that dataset. Hence, as has been pointed out in many of the discussions sparked by Open Life Science, the current paradigm of open science has systemic power and privilege asymmetry built-in. The primary barrier is the lack of candid conversations that end up in accountable actions to change the paradigm and dismantle the asymmetries. Until asymmetries are gone, we must change the system metrics. We need recognition for collecting and maintaining data, writing and maintaining code, publishing a paper, or organizing welcoming communities and events. These activities are crucial for open source and science to thrive, yet only publishing the paper counts in academia. Also, I think we must give extra credit if people do any of these despite systemic marginalization playing against them.
Fernando: Open source has created new and complex relationships between academia and industry. How do you see the culture of open source and open science in your countries’ academic circles?
Laura: In Latin America, we are where the U.S. was a decade ago in terms of adoption, participation, contribution, and leadership in open science. Inclusive communities of practice that lower as many barriers as possible to international involvement, such as R-Ladies, PyLadies, Open Life Science, or The Turing Way, help to change this. But we need many more. Also, communities of practice will be a lot more effective if they start at least favored geographies and get to have international reach, rather than starting in high-income countries and getting global reach from there.
Gonzalo: Latin America uses a lot of open source tools, but we still need to make the jump to contribute to and create them. We need more code sprints, more visible faces, and more incentives for people to contribute. We have to keep in mind that a lot of open source contributions are enabled by the “free-time” of volunteers and this is a luxury that many cannot afford. To be able to really make progress, we need academic incentives for students and also for professors.
Fernando: As more companies from Latin America expand their reach, what could they do to support open source contributions? How could ties between academia and these companies improve?
Gonzalo: These companies are very big users of open source technologies and in some cases, also creators of open source software. However, we are still struggling to have these big Latin American-based companies contribute in a meaningful way to open source. Individuals in these companies at some of the more technical levels understand the value in contributing to open source, but this needs to go up into the C-level and managerial levels so a true commitment to invest in open source is made.
To improve ties between academia and these companies, there should be a lot of opportunities for students to do internships in technology companies working and using open source software. A point of intersection is the work with open source communities, so I see a lot of value in driving more academics into the open source communities where industry and academia can meet. Academia could also host more events where industry takes an active role, such as tech fairs and science fairs.
I have seen many cases of people from academia that moved to industry. I stopped my PhD because I was not happy with what I was doing, but also I always wondered what I would do when I returned to Colombia. Do academics have the opportunities and resources they need? The majority do not. PhDs have to look for different opportunities, and being trained scientists, migrating to be a data scientist is an option. Not that these PhDs do not enjoy their new life as data scientists, but many landed there because there was no other choice. I have also seen PhD graduates migrate to software engineering.
Fernando: What are concrete ways in which we can make open source and science genuinely global?
Gonzalo: For traditional education programs, we need all curricula to include an introduction to programming. Much like English is a need, programming literacy is also needed. We also need bilingual education from primary school all the way up and reliable infrastructure. Before we can make things global, and we reach full English literacy, we need to make it local. We also need to strengthen the ties between academia, industry and open source communities and advocacy groups.
Laura: Many experts in open source and open science in Latin America are open to collaborating. Reach out! Let’s talk! Collaborations could include applying for grants or sponsorships together, contracting jobs to develop training materials, consulting services from experts in the territory, or evaluating if what you think is inclusive for Latin Americans is actually inclusive. The mantra of people with disabilities of “nothing about us without us” works perfectly for Latin American scientists, too. Funding more incubators such as CS&S that give access to funds usually unavailable internationally and provide all the infrastructure to execute them. Fund more talent-leveling organizations such as Open Life Science or MetaDocencia.
We need more on-ramps programs teaching open leadership, community building, and working openly in open science. Pay for the time of local folks to learn open science and open source. Make learning materials accessible by having them in local languages. If you are overrepresented, make room by offering your seat at decision-making tables to marginalized folks. You can also condition your participation to marginalized folks joining and staying at the table. Always assume you have a lot to learn from others. Last but not least, if you want to have a diverse global audience in your event, lower the economic barrier. Always offer a free ticket option, no questions asked.
A closing thought
Our world is highly interconnected, but while a tweet can be read instantly across continents, active engagement in the creation of the computational infrastructure that underpins science remains unevenly distributed. For the authors, the key promise of open source science is not access to tools or knowledge free of charge, but participation in the creative process that is open to all. While the basic tenets of scientific discovery transcend boundaries, local context is critical when it comes to actually understanding the world we live in. Our countries face challenges where the tools of global science are of critical importance, yet we need those tools to adequately represent our reality. We can create that knowledge if we have equal agency in shaping the tools of science itself.
Laura Ación is a research scientist at the National Research Council in Argentina, where she leads a multidisciplinary artificial intelligence and data science co-laboratory. With a Ph.D. in biostatistics from the University of Iowa, she works on topics related to the responsible use of data. She is an open science community builder, who, for example, co-founded and leads MetaDocencia, a Spanish-speaking community for open science capacity building.
Gonzalo Peña Castellanos is a Colombian Civil Engineer with an MSc. in Hydroinformatics and an MSc. in Sanitary Engineering that currently works with Quansight, Inc as a Senior Software Engineer. He is an active member of the open source community contributing to different projects as a member of the grants working group in the Python Software Foundation and a leader in the Python Colombia Community helping shape and organize local and regional communities.
Fernando Pérez is faculty at UC Berkeley Statistics and a scientist at LBNL. Trained in physics, he builds tools for humans to use computers as companions in thinking and collaboration, mostly in the scientific Python ecosystem. Today, he focuses on open, reproducible science at scale to tackle problems like the climate crisis that bridge physical modeling, data analysis and societal concerns. He co-founded Project Jupyter, 2i2c.org and NumFOCUS.