Virtual Panel: How Open-Source is Helping to Change the World

Open-source is not only at the heart of the Cloud and the enterprise as we know them today, it also powers many initiatives around the globe that aim to change how technology is leveraged to solve real-world issues and strive to make it accessible in a more equitable way. InfoQ has taken the chance to speak with several companies that are helping make it happen.

There are many initiatives and projects for social good that include open-source software among their tools. Their goals are diverse and span from reducing domestic and gender-based violence to empowering the LGBT+ community through the tech sector, from helping other nonprofit organizations to fight racial discrimination in technology.

Two noteworthy projects are Global Forest Watch, a platform providing data and tools to better protect the forest, and OpenAQ, an organization aiming to empower communities towards clean air and air equality.

Global Forest Watch created a platform that gives access to near-real time information about forests around the world, allowing to monitor their state and change, including deforestation, fires, and so on. OpenAQ has built a community of scientists and software developers to create an open, real-time database for air quality data.

InfoQ has spoken with Mala Kumar, Director of Tech for Social Good at GitHub; Thomas Maschler, Lead Software Engineer at Global Forest Watch; and Sruti Modekurty, Platform Lead at OpenAQ, to learn more about the challenges and opportunities of using open source software for social good.

InfoQ: Open source and community were the two driving forces of a large transformation in the software industry that redefined the way we build software and systems today, away from closed spaces and protected silos. Could we say those same factors are now driving the creation of a new approach that will redefine how our society tackles its challenges at large?

Kumar: I think the answer depends on the industry or thematic area on which you’re focusing. Open collaboration in public health has been around for decades, which was discussed in a presentation with the World Health Organization last year. That’s how diseases like polio have been (nearly) eradicated globally and how global scientific communities came together to create mRNA vaccines for COVID. It’s not surprising with such a natural alignment, that a lot of our collaboration around open source in the social sector is in public health.

I do want to caveat, however, that digital technologies - including open source - play to both sides when it comes to human development and poverty reduction. On one hand, the past decade has seen massive strides in technological advancements, yet many of the same geographies that are at the forefront of those advancements have also seen rising poverty and income inequality.

Open source may be open, but open does not always mean equal. Who can access the technologies and participate in a safe way is not the same across all demographics, geographies and populations. The social sector tends to be relatively advanced and conscious of these areas, and so one of the great effects we’ve seen through our work is bringing the best of both worlds together. Imagine a tech project for the advancement of social good that uses open source tools and methods and does so in a racially, gender and socioeconomically equitable way? That’s powerful!

InfoQ: What are the biggest technological challenges in projects like Global Forest Watch and OpenAQ?

Maschler: The biggest technological challenges for our projects remain with processing the raw remote sensed data. The global forests are highly diverse and can appear very differently depending on the biome and season. Our partners at the University of Maryland and the University of Wageningen already did a fantastic job in reliably detecting forest loss at global scale in such a high resolution.

We currently still miss globally consistent data on forest gain at a similar temporal resolution as our forest loss data. It takes much longer for forests to grow than to cut them down. The smaller the change, the harder it is to detect - especially from space.

Our partners at the University of Maryland are currently working on an exciting project to measure annual forest net change based on tree height. They calibrate optical images from the Landsat mission using laser based measurements from Gedi and IceSat-2. Once these data sets are released we will get a much better picture on forest dynamics and underlying drivers.

Modekurty: One of the biggest technical challenges is maintaining and scaling the open data infrastructure since we aggregate and harmonize air quality data from 100+ disparate data sources around the world. The sources may become temporarily unavailable, the data format may change, the coordinates can change slightly, so the platform needs to be robust and modular enough to work around unexpected changes.

The platform currently hosts 4+ billion data points from 25k stations in 120 countries and as we increase data coverage and temporal granularity we need to scale up data storage and capabilities while remaining cost-effective as a small nonprofit!

The other main challenge is effective design and tech literacy. We want to make sure the open source tools and toolkits we are designing and developing are accessible to different audiences with varying degrees of technical expertise and continue to support a number of different use cases and users, from researchers and developers to activists, journalists, communities and individual citizen scientists.

Being open source means we can rely on our awesome global open source community to help us fix issues when they arise, add new data sources, and develop additional tools and educational materials.

InfoQ: Collaboration is at the heart of many open-source projects to bring awareness and to build solutions to a number of problems of our global society. From the vantage point on collaboration and open-source projects that working at GitHub provides, what are the other key ingredients of successful open-source community projects?

Kumar: Above all, when you’re building tech for social good, there’s a level of context building that is absolutely essential. We have a saying in international development called "do no harm." People working in the social sector can’t take the same kind of risks that strictly corporate or for-profit ventures may take, because if something goes wrong, we could put vulnerable populations in danger, waste the few assets or time they have, build expectations and let them down, and basically make their lives worse. The goal is not about finding a new market opportunity (i.e. customer base), it’s about ensuring the tech makes lives better.

For that reason, any tech for social good community (including open source), must have that context building component. That may mean the community is attached to an organization, company, government or other entity that has the power to make real change. Context building can also come from members with a social sector background, or that the work involves people who are of the population the tech community is trying to help.

InfoQ: Based on your experience, what is the force that brings together and makes successful a community-driven project for social good?

Maschler: Unlike for-profit websites where ultimately revenues are the decisive factor for which feature to build and which to drop, we as a non-profit have to invest much more in user engagement and outreach as well as monitoring and evaluation to get it right.

From the beginning, Global Forest Watch was designed as a partnership of many different actors. From big tech firms like Google and global commodity traders like Cargill, to national forest agencies like OSINFOR, news networks like Mongabay and local NGO like Forest Defenders, everyone has something to contribute. We host regular events where we bring all partners together and have them share their outcomes and challenges and discuss future priorities. Our engagement team works with national governments to adopt our data and tools, and builds a network of NGOs and individuals which can act as multiplicators. Our small grants fund program allows local NGOs to adopt our tools and data to their local context or come up with new innovative ideas.

Our Monitoring and Evaluation team keeps track of success and failures and helps us to stay focused on making our tools as impactful as possible.

Modekurty: We believe open data is a key driver in bringing together different stakeholders because it creates a level of transparency and starts to get to a shared understanding, breaking down silos in the air quality ecosystem.

OpenAQ believes that stakeholders convening around AQ data (in person and virtually) are key to developing the solutions required to address air inequality, the unequal access to clean air. While data is an important starting point, the solutions that come from that data still need to be participatory and co-owned by different stakeholders, and most especially, to be inclusive of the voices of marginalized communities often most affected by air pollution.

If our goal is to democratize access to air quality data and information and support action, we also need to make sure we are bringing varied technical, thematic knowledge, and lived experiences to the table.

InfoQ: Could we talk about how local issues prevalent in economically developing regions are being addressed by adopting low-cost technologies, particularly in ways that can be replicated in different contexts or applied elsewhere to different issues?

Kumar: That is a massive question and there are tens of thousands of people globally working on different aspects. A good way to learn more is searching for the term ICT4D, which stands for technology for international development. I gave a talk in Belgium last year about some examples in UX design. Back in January, I wrote about an interesting example of open source contributions from the social sector from the company Dimagi, which runs the CommCare app.

Maschler: Access to internet and computer literacy in the "global South" remain a challenge. Only about a quarter of our website users come from South and Central America, Africa and South-East Asia; the regions where most of the loss in primary forest happens. Our engagement team is working closely with organizations in these priority regions to better understand their needs and challenges and to identify solutions which work even under challenging conditions.

Our mobile app Forest Watcher was designed together with law enforcement agencies in Uganda and was built to work without permanent internet connectivity. It can run on simple mobile devices.

We allow users to subscribe to our alert system and receive notifications once we detect change in the area they are monitoring. This way, users do not have to constantly log on to our website and consume expensive internet credits in order to check for updates.

We also invested heavily in training key players in using our tools through our regional offices. In Central Africa, where we have a strong regional presence, we could detect an 18% drop in deforestation for areas of active monitoring.

Every year we give out small grant funds to local NGOs. This offers applicants the possibility to try out their own ideas, adopt technologies their way and directly engage with local communities which live closest to the forest. Many of the lessons learned from these projects flow back into our product cycle and help us to improve our tools and make them work for all our users.

I believe that we found a good approach to serve out zonal statistics at scale. Calculating area statistics using high resolution data is very compute intensive and serving out those numbers at webspeed remains a challenge. The tools we build to do so can be easily adopted for any other spatial analysis outside the forest domain (e.g. cities, finance, agriculture, telecommunication, and advertisement).

Modekurty: In the air quality arena, OpenAQ’s Global State of Play report (released summer 2020), identified that while more than 90% of the world’s population experiences unhealthy levels of air pollution, only 50% of the world’s governments produce air quality data, leaving over 1 billion people without access to air quality information in their country.

Government-sponsored, reference grade monitors provide a key and important standard for driving local air quality efforts, while newer low-cost sensor technologies can help fill coverage gaps, and provide opportunities for communities, citizen scientists (and also governments) to produce additional localized air quality data that contributes to the fight against air pollution.

Recognizing this opportunity, OpenAQ recently added low-cost sensor data to our platform through a pilot project, in partnership with Environmental Defense Fund and three additional low-cost sensor partners, something we will be expanding upon in the coming months. While there is still a long way to go to get sensors into the communities that are most affected by air pollution, lower cost technologies certainly can help lower the barrier for entry. At the same time, as with any emergent technology, there is always a question about cost and quality and what it means to scale these efforts.

To address the scale question that new technologies present, we are also asking what it means to scale training efforts more widely in order to use AQ data to drive change, as well as the type of information that communities need to best access and use this data. To this end, OpenAQ has recently launched a regional model through our newly formed Community Ambassador Program (a 10-month fellowship with a "train the trainer" focus to support global air quality advocates). Through this program and other opportunities for engagement with our stakeholders, we are developing key community feedback loops in order to best develop and compile the tools, toolkits and information required to ensure that both existing and emerging technologies (and the data they produce) can be well utilized by those who need them most. We will be looking to expand the Ambassador program in the coming years, and also look forward to sharing lessons learned with others exploring innovative ways to scale local change efforts.

InfoQ: I am sure this will be hard to say, but if you had to choose one social good-oriented, community-driven, open source project that had the most impact in the past, which would you name? And the one that could have the most impact in the near future?

Kumar: That is really hard to say! There are obviously many amazing tools and by no means do I wish to imply any are more important than others. One open source for social good tool that has truly transformational effects is DHIS2, which is the largest health management information system (HMIS) in the world. When I was working on a project for USAID, I used DHIS2 quite a bit in Kenya. When I was working for UNICEF, my colleagues at another organization led the first DHIS2 rollout in Burundi. DHIS2 has continued to expand its reach, and there are now instances deployed in 70+ low and middle-income countries.

For resource-poor countries or sub-national geographies, DHIS2 is a critical piece of public health infrastructure. The DHIS2 developer community is also on GitHub, and we’re really proud that they’ve chosen to make GitHub its home!

About the Panelists

Mala Kumar, Director of Tech for Social Good at GitHub is an experienced tech for social good practitioner with a strong emphasis on UX research and design. She has designed, led and collaborated on ICT4D projects throughout sub-Saharan Africa, and in NYC, DC, and Europe in English and in French. Mala founded and currently runs the Tech for Social Good portfolio of programs at GitHub. Her previous experience extends to the United Nations, the private sector and INGOs.

Sruti Modekurty, Platform Lead at OpenAQ is passionate about using tech to solve problems related to environmental change and civic engagement with a focus on equity and justice. She has a BSc in Electrical & Computer Engineering from Carnegie Mellon and has a diverse engineering background spanning industry, government and nonprofit sectors. She is currently on the board of hackNY, helping talented CS students from underrepresented backgrounds use their skills for good. Previously, she was a fellow at NYC Mayor's Office for Economic Opportunity and Blue Ridge Labs, working on tech-enabled solutions for New Yorkers facing housing instability.

Thomas Maschler, Lead Software Engineer at Global Forest Watch leads the Engineering team of Global Forest Watch (GFW) and is responsible for providing technical leadership and vision for GFW’s data management and architecture, which are the cornerstones of the GFW Platform. Thomas holds a Diploma in Forest Ecosystem Management (German master's degree/M.Sc. equivalent) from Albert-Ludwig-University Freiburg, Germany. For many year he has worked on forest cover change and land use conflicts related projects using GIS and remote sensing technology. He also served as technical lead and product manager for various WRI developed applications. His former employers include Vodafone, Worldbank, and GIZ. He enjoys outdoor activities, biking and keeping livestock.