The Homebrew Sensing Project: Environmental health monitoring with open source tools
This is the future we imagine through the Homebrew Sensing Project. A dizzying array of hazardous chemicals surrounds us—from formaldehyde in building materials to brominated fire retardants and dyes in the clothes we wear, to additives in food and industrial contaminants from hydrofracking. People increasingly worry about the health impacts of their exposure to these chemicals on their bodies, families and communities. As whole communities face chemical exposure and related health problems at a large scale, these issues become a concern we share with our neighbors, debate at town council meetings, and join forces to solve through action and advocacy.
Public awareness of environmental toxics and their effects is becoming a daily part of our lives as people look for chemical-free food options in organic produce, search labels on plastic products for BPA, and shop for low-polluting, high mileage cars. The reality is that we all come into contact with toxics several times each day as we travel, work, shop, and play with our children. Many of these chemicals are linked to birth defects, respiratory problems and cancer, among other health problems. Even today, the impacts are especially felt in Environmental Justice communities—often rural or low-income communities within sight of industrial facilities—who are subjected not just to the normal barrage of toxics but to a disproportionate amount of pollution and runoff from industrial agriculture, manufacturing and petrochemical refining. In these communities, breathing in chemicals like benzene, volatile organic compounds and hydrogen sulfide is a part of daily life, resulting in both short- and long-term effects such as heightened risks of cancer, asthma, neurological disorders, rashes, reproductive disorders, and more . Due to limited reporting requirements for industry, residents have very little information on what they are breathing, the concentrations of the toxics, symptoms, and helpful preventative behaviors.
Currently, people concerned about chemical exposure must rely on the "send it to a lab" paradigm—where residents or advocacy groups collect air or water samples for off-site analysis by experts—an expensive and cumbersome process. Lab science has always been beyond the reach of families and everyday people, requiring capital equipment, training, and formalized expertise. Basic lab-based air quality analysis, for example, can cost $500 per sample (on the low end) and has strict time constraints on when it must reach the lab. But what if people could collect and analyze samples in the field, in real time, for 1/100th the cost?
Just as the Homebrew Computer Club challenged assumptions that computers had no place in the home, the Homebrew Sensing Project explores the potential for spectrometry, a powerful analytic technique for identifying materials, in your home and neighborhood. We have leveraged the emerging Do-It-Yourself movement and the power of open source hardware and software development to bring spectrometric analysis to a new audience—non-experts investigating and testing their immediate environment. The Homebrew Sensing Project gives everyone access to simple, affordable tools to collect data in their own backyards, providing the information they need to make environmental health decisions.
This project will draw upon the knowledge of over 700 active members who are already working together with Public Lab to bring low-cost chemical analysis to everyday people through a collaborative, open source development process; 1,600 spectral data contributors, and over 3,500 people using and building prototype tools globally. Our open spectrometry community brings together people from all corners of the globe and all levels of knowledge—local environmental health advocates in refinery communities of the Gulf Coast, retired professional scientists looking to contribute their skills, and elementary school students inspired by hands-on learning—to improve the hardware and software, test tools in the field, compare results, develop research questions, and design experiments. Prototype spectrometers being used by the community include a foldable, paperboard spectrometer that refracts light with a piece of DVD-r, a molded plastic spectrometer that mounts on the back of a smartphone, and a robust desktop version that can be built with basic supplies from a hardware store. The collection of hardware is supported by Spectral Workbench, open source spectral analysis software developed by the Public Lab community. Using Spectral Workbench, contributor samples are aggregated into an open-source library of spectra; data is collected, compared, and analyzed; and discussions about technique, refinement, and use-cases occur among the global community of users.
Early successes include basic detection of sodium and mercury, as well as promising results from measuring the fluorescence of petroleum-based samples using a laser or UV light. At the moment, community members are using the power of the Homebrew Sensing Project in a variety of ways:
* Identifying petroleum in the sediments on Grand Isle, Louisiana with a team of coastal ecologists and activists exploring the long term impacts of the BP oil disaster.
* Developing techniques to monitor emissions flares from a distance at refineries in the Gulf Coast, in partnership with NASA DEVELOP and the Gulf Restoration Network.
* Determining the presence of chemical coloring agents in detergents and household cleaners by parents worried about the products they use for their families.
* Teaching chemical analysis to students, from elementary school through college, using hands-on tools and locally relevant samples, like motor oil runoff on city streets.
The Homebrew Sensing Project empowers people around the world to engage in projects like these, harnessing the power of local, grassroots “Small Data” for the health of their communities. Big Data—the idea that the ability to aggregate and sift through vast amounts of data can yield key insights about our society and provide the basis for better decision-making—though increasingly popular, has a fatal flaw. It is premised on an asymmetric, purely upward, flow of data towards a central authority whom we must trust to make decisions on our behalf. In this age of participation, Do-It-Yourself, and civic empowerment, shouldn't we all have a seat at the table?
The Homebrew Sensing Project supports the creation of technologies which enable affordable, accessible data production, collation, management, visualization and analysis—so let's leverage it in the name of Small Data; a bottom-up, voluntary, shared model of data collection and aggregation whose participants are not mere data points. Small Data ecosystems will be built on the open exchange of data, by and for the public, towards civic ends. Small Data goes beyond crowdsourcing; it is based not merely on public contribution to a common goal, but on the public having a say in how that pooled data is used, and what questions it answers.
With the support of the Knight Foundation through the Health Challenge, the Homebrew Sensing Project will grow both in scale and scope, improving the hardware, software and collaborative interface citizens are already using to collect data for health advocacy in their communities. The project will support improvement and revision of the hardware tools based on community research and input, as well as improved features of Spectral Workbench, including advanced capabilities for matching unknown samples against known spectra (building on work started this summer in partnership with Google Summer of Code), a mobile-friendly application and an offline interface for easier use outdoors and in remote areas—the latter two being key to making the tool more accessible and field-ready. Other planned improvements include better collection and management of positive and negative control data, ease-of-use refinements, an expanded API and scripting language, and compatibility with government chemical databases.
The Homebrew Sensing Project will expand outreach efforts to reach an increasingly diverse and inclusive audience, with special attention given to Environmental Justice communities along the Gulf Coast. We will encourage collaboration among local users by hosting regular sensing meetups in key locations, as well as hosting a global competition, the Spectral Challenge, to prompt teams to work together to improve tools and complete verifiable toxics identification in areas of concern in their communities. Knight Foundation support will also provide for the creation of a Community Health Toolkit, which will empower users with comprehensive information on data collection, analysis and aggregation, and will also be instructive on taking the data to action, guiding communities on how to best use their data to advocate for improved environmental health.
 See for instance:
a. Environmental Integrity Project. “Houston, We Have a Problem: A Roadmap for Reducing Petrochemical Industry Toxic Emissions in the Lone Star State.” www.environmentalintegrity.org. 2008. http://www.environmentalintegrity.org/pdf/publications/Houston_We_Have_A_Problem.pdf
b. Center for Environmental Health. “Toxic and Dirty Secrets: The truth about fracking and your family’s health.” www.ceh.org. 2013. http://www.ceh.org/legacy/storage/documents/Fracking/fracking_final-low-1.pdf
c. Environmental Protection Agency. “Fast Facts on Children’s Environmental Health.” www.epa.gov. 2008. http://yosemite.epa.gov/ochp/ochpweb.nsf/content/fastfacts.htm
Who is working on the project? Who are your partners?
This project is led by our global community of contributors--the people who are actively engaged in using, improving and sharing the hardware and software tools that make up the Homebrew Sensing Project. Local community members provide contextual expertise and are most capable of taking these tools into the places they care about to measure and track toxics and health impacts. The community will receive support and guidance from a team of software and hardware developers, community organizers, and environmental health advocates at the Public Laboratory for Open Technology and Science, including Jeff Warren, Shannon Dosemagen and Mathew Lippincott. Public Lab staff will provide technical expertise and infrastructure to ensure that the project and tools are moving forward, assist with outreach to grow the community in an inclusive, diverse manner—especially in the Gulf Coast—and will facilitate partnerships that require institutional support, including supporting and expanding existing relationships with NASA DEVELOP, the Gulf Monitoring Consortium, and regional community organizations.
How do you know there is demand for this project?
When we launched the Homebrew Sensing Project pilot via Kickstarter in Fall of 2012, we garnered interest from over 1,600 people who contributed nearly 500 different ideas on how they could use the spectrometer to look at things such as water, air and soil quality. The launch of this project resulted in the growth of what had formerly been a small but dedicated Public Lab spectrometry community into an active, engaged, diverse group of practitioners interested in developing new modifications on the open hardware design, educational curriculum for teaching spectrometry, “in the field” workshops, and new research directions. Over the past year, over 1,500 ideas and comments have been shared on the Public Lab spectrometry discussion group by 700 active members, and over 1,600 contributors have shared in excess of 10,000 spectra since our open data repository launched ( http://spectralworkbench.org/stats). What is more, we have distributed over 3,000 spectrometers through our Starter Kits initiative in less than a year, in part due to our emphasis on affordable kits priced at $10 and $40—including a prototype smartphone attachment—and we have committed to developing mobile applications to make this kind of sensing more accessible and portable.
How is your project different from what already exists?
Unlike projects which are content to “just visualize the data”, assuming this alone will effect change, the Homebrew Sensing Project tackles the process of data collection itself—reducing the cost of analysis tools by a thousandfold, and developing techniques that can be used, understood and modified by novices to generate rich, open source, online chemical data. We’ve worked to challenge a model where knowledge is created solely by accredited, well-funded institutional scientists and distributed to the public primarily as a form of outreach, instead placing scientific power and control over data collection and analysis in the hands of people who need it most, to advocate for local environmental health issues. Unlike other citizen science projects, we aren’t simply deploying people to collect data points for a centralized research project--our community engages in hardware and software development, frames research around their own concerns and questions, shares knowledge and solves problems collaboratively. This isn’t just a hardware project, an open software project, or a citizen science initiative—the Homebrew Sensing Project leverages the power of each of these models in building a fast-growing, globally available, easy to use, community-supported chemical analysis platform to help people answer questions about the health of their homes and communities.
How will the data or information you use or create be made open?
As a project which integrates open hardware, open source software, and open data, the Homebrew Sensing Project embraces openness throughout our process—we share designs for the devices, code for collecting, analyzing, and sharing data, and run an open data repository to crowdsource chemical identification. Using the CERN Open Hardware License, the GNU General Public License, and the Creative Commons Zero license for hardware, software, and data, respectively, we have built a collaborative network of thousands of contributors across all three domains. Furthermore, unlike many “vacuum cleaner” data repositories, our Spectral Workbench website allows users to create and share a body of their own research, while commenting, “liking” and tagging bring the data to life by cultivating discussion, debate, and peer learning. The software not only helps with analysis, but becomes a social and technical platform for building new tools and finding collaborators.
What will you make or do in this project?
The Homebrew Sensing Project is part of a larger, long-term effort by Public Lab and its network of contributors to reimagine environmental health science as something the public can and should make use of in everyday life—using a combination of personal inquiry, basic science applications and local environmental health knowledge to democratize the scientific process. This project is an important step towards that goal because it will energize a global network of civic science practitioners, innovating, documenting, communicating and advocating around similar but distributed problems related to environmental health threats and their impacts. This knowledge base will be used to create a Community Health Toolkit—a guide for other communities to use these tools to advocate for local change, including hardware and software, tutorials, and best practices for advocacy. This project will enable us to expand and improve upon our progress to date, upgrading and improving our software, providing for mobile-optimized interfaces and designing for offline “fieldwork” mode and for 24-hour monitoring abilities.
How can others learn from/build on what you do?
The Homebrew Sensing Project, as a Public Lab initiative, has been well documented from the start through regular, openly published research notes, wiki pages and tutorials—all of which play an important part in our collaborative research methodology. The open licenses we use for hardware, software, and data have already proven key to allowing imaginative reuse and adaptation of our work by new contributors and independent projects alike; for example, several other open scientific data projects have reused or cited inspiration from our open source Spectral Workbench software. A rich API and scripting language have allowed developers and hackathon participants to easily scrape, remix, and build upon our data, and the Public Lab web platform itself has also attracted new open source science and environmental health projects as a way to tap into the vibrant Public Lab network and attract interest and collaborators.
How much do you think it will cost?
This project will cost $350,000 over two years, including staff time for hardware development; software and mobile app development; outreach and training; and advocacy. Additionally, funds will cover tool prototyping expenses, including stipends for community partners field testing the hardware in the Gulf Coast, as well as program outcomes monitoring and evaluation.
How would you use News Challenge funds?
News Challenge funds will be used to support two phases of work for the Homebrew Sensing Project, with each phase lasting 12 months. Phase One will focus on development and refinement of hardware and software tools, including hardware development time to refine the spectrometers and instructions to ensure easy production and clear data, and software developer time to refine Spectral Workbench, build out an offline version for field settings where internet is unavailable, and create a mobile-friendly version for regular field use. Phase Two will focus on implementation and outreach with an emphasis on the Gulf Coast, building on existing relationships with nonprofits, technical partners and community organizations to distribute tools to users and host local and online trainings and meetups. Additionally, Phase Two will include final development, based on user input and feedback, of the Community Health Toolkit, which will include start to finish guidance on tool implementation and advocacy.