Drip by Drip: Creating a Water Commons

Photo by Code for Africa

Despite its overwhelming importance, hard data on water access and quality is hard to come by. A small band of journalists and civic tech groups are pooling resources to fix the problem.

In August 2020, Code for Africa, OpenUp, Code for Germany and the Center for Collaborative Investigative Journalism teamed up to collaboratively build a repository for global water data. All five of our groups had been working on water-related projects and were encountering similar problems locating data and putting it to work. What a shock, since water is vital to everything we do as human beings. The United Nations has even declared access to clean water a universal human right, and its Sustainable Development Goals rightfully take note of the critical role data plays in securing those rights.

Yet, here we were, facing a kind of information drought.

Enter the Code for All Exchange Program, funded by the National Democratic Institute and the National Endowment for Democracy. The program gave us a chance to pool our technical and investigative resources for a much larger aim: to support those who lack access to clean water. In particular, we wanted to build something that would empower journalists and civil society actors to press forward in this quest; as their storytelling and research into local water grievances can help activate communities for change. It was also important for us to test and leverage the expertise of the Code for All Network to overcome the data and resource barriers we were facing in our separate projects.

Building on previous projects

Though we never established clear cut roles, our organic collaborative style seems to be working thanks to the skill and commitment everyone brings to the table. There is OpenUp that developed Wazimap, a spatial open data portal to make geographically-linked information more readily available to non-technical audiences. A lot of rich demographic data and even some water-related information had already been collected into the platform’s central database, making it a snap for our tech team to fire up a new “instance” focused primarily on the new data we’ve been collecting for the Water Commons itself.

We greatly profited from the insights into journalistic needs and approaches provided by the Centre for Collaborative Investigative Journalism (CCIJ), a non-profit that was already producing several data-driven and visually rich investigations into water rights.

Code for Africa, an impact accelerator that supports a range of water focused data citizen science initiatives, has been working on water quality testing kits for water-borne diseases. Additionally, CfAfrica’s forensic data science team, the iLAB, helps transform this data into actionable insights for a network of watchdog CSO and media partners, performing the interface between both worlds.

And Code for Germany has dealt with water data on a more administrative level, trying to convince decision makers to open water data to the public. To demonstrate how this data can benefit the population overall, CfG developed a drinking water app that visualizes the quality of drinking water in municipalities throughout Germany.

First steps and problems encountered

We set up a weekly Jour Fixe (regular group meeting), which helped us gel as a team and feel comfortable discussing project issues as a collective. These standing meetings also helped us avoid most of the scheduling headaches that might otherwise plague a team of Africans, Europeans and Americans; though, as a kind of professional exchange bonus, we did learn about one another’s holidays, pets, snowstorms and even some construction work that occasionally made virtual attendance complicated. More beneficial, our weekly meetings have helped us minimize the duplication of efforts.

This was also part of our motivation when we asked Code for All members and others in our broader community to share any water-related projects they might have underway. Several colleagues pointed us to water data in their home countries and regions, and their responses in general underscored the importance of water to the civic tech community. However, existing project commitments and COVID-19 seemed to conspire against us as we tried to rope in more collaborators, especially when our work was just getting started and our ideas yet to take concrete form.

During these early days, we put a lot of effort into researching water data and hunting down any existing tools. No need to reinvent the wheel. But it became clear that we were in fact building something useful, especially if we could combine data from many countries and subnational levels in one place.

Image for post
Water samples taken in Makoko [Photo by Code for Africa]

Lessons learned

At this point it’s quite clear that there really isn’t enough water data available and that, yes, it is hard to find! Although international actors like WHO and UNICEF provide important global baseline information, subnational data is particularly scarce and rarely standardized. And water testing is still mostly a governmental domain, leaving few alternatives.

Simply put, gathering water quality data is an elaborate and expensive process. Because of the corrosive effects of water, you can’t leave toolkits submerged and unsupervised for long periods of time. Depending on the type of water body, samples have to be taken at different depths and locations and sample-takers have to be trained with the equipment.

With this problem in mind, we spent a significant amount of time searching for more basic assessments of water quality: photos for visible attributes, characterization of smell or even the presence of animals and microbes. We stumbled across the miniSASS project early on as a promising lead, which simply asks citizen scientists to collect samples of macro-invertebrates in any body of water to get a general measure of water quality. This low-tech, inexpensive approach piqued our imaginations, and when we suggested a potential partnership, the miniSASS developers showed mutual interest.

But even if we could measure water quality cheaply and uniformly, we uncovered other vexing problems. One of the biggest challenges facing journalists is simply a lack of localized data with the specificity they require. Water data is often presented at the aggregate level nationally, making it difficult to pinpoint problems within specific communities. And when we asked journalists in our network to describe the data they were after, they identified a dizzyingly diverse collection of needs: borehole locations in Nairobi; dam efficiency and water-level indicators; water pipe replacement and maintenance schedules; the cost, efficacy and byproducts of specific water treatment chemicals; and much more.

In Lagos, Nigeria, CfAfrica has been hard at work in the Makoko community tackling this information shortfall head on. The team trained groups of volunteers to take samples in local water sources and worked with local researchers to produce detailed lab results; all of which we are loading into our Water Commons database. The effort has given our team a good sense for how expensive and labor intensive water testing can be, but we’ve also seen an exciting level of engagement at the community level.

Image for post
Municipal Council meeting with the water volunteers
Image for post
The water volunteers with our samples [Photo by Code for Africa]

Meanwhile, CCIJ put out a call to its network of journalists asking them to uncover interesting datasets and pitch powerful accompanying stories. This approach ultimately forced us to think very broadly about what might constitute Water Commons data. Among the projects CCIJ decided to support: A photographic examination and data visualization of the environmental causes of land subsidence (literally sinking land) and its impact on one family in Gouda, Netherlands. In another story, CCIJ teamed up with a group of journalists in South Africa that has been scraping data from municipal websites in a bid to better understand the plight of the country’s overwhelmed wastewater infrastructure.

Working directly with journalists has generated a number of weighty questions. Do we include huge datasets, such as the satellite imagery being used to tell the land subsidence story in Gouda? What about other climate change data or information on health care institutions and labs? Would journalists want to measure a community’s access to drinking water by measuring, say, average distance to taps?

Journalists have also told us they want support gathering, cleaning, analyzing and visualizing data. How can we configure the Water Commons to meet those needs? We are now mulling over the need for complimentary online courseware that can offer assistance, along with the many visualization tools we hope to bake into the data repository itself.

What’s next?

Currently, we are in the midst of a massive effort to clean several of the datasets we’ve uncovered during our research and to upload them into a working prototype using the Wazi platform. Datasets must be transformed into collections of specific indicators that are tied to a geographic area; topics, country names and data formats are also appended. In this way, we are distinguishing between global data (more than three countries) and national data (fewer than three countries). Due to our broadened search, we have been able to cover a wide range of water rights’ indicators, from water quality measures, dam levels, water infrastructure assessments and more. But, as one might expect, we are locating data in a plethora of formats, from CSVs and APIs to PDFs and simple web interfaces that require scraping; and the data preparation work has proven to be quite time consuming.

Not only do we need to complete more of this work to make our prototype usable, but the team will need a good plan for how to incorporate more data down the line. And we need to tackle “front-end design” challenges going forward, including a better way to categorize and organize the data for Water Commons users.

Image for post
Water Commons logo
Image for post
Water Commons prototype interface

And we know there are other interesting ideas we’ll need to discuss; for example, we’re considering the value of creating and adding a virtual Rolodex, if you will, to the Water Commons to help journalists find credible sources and local water activists. Or, if we partner with the miniSASS creators, will we need to modify our work to better target groups like youth in general or Fridays for Future activists. In short, we’ll need to adapt the Water Commons to potentially meet the needs of a very broad audience.

While our Water Commons beta instance is already online, it’ll be clear to visitors that it’s truly a work in progress. With many of the challenges unresolved, ideally, we want to start another round of tests and collaboration with journalists, now with a more substantial tool already at place. As with most projects like this, more funding is needed. But if we get it right, we can build a tool — collaboratively and drip by drip — that helps fulfill the right to water.

Share this post:
Email icon

Despite its overwhelming importance, hard data on water access and quality is hard to come by. A small band of journalists and civic tech groups are pooling resources to fix the problem.

In August 2020, Code for Africa, OpenUp, Code for Germany and the Center for Collaborative Investigative Journalism teamed up to collaboratively build a repository for global water data. All five of our groups had been working on water-related projects and were encountering similar problems locating data and putting it to work. What a shock, since water is vital to everything we do as human beings. The United Nations has even declared access to clean water a universal human right, and its Sustainable Development Goals rightfully take note of the critical role data plays in securing those rights.

Yet, here we were, facing a kind of information drought.

Enter the Code for All Exchange Program, funded by the National Democratic Institute and the National Endowment for Democracy. The program gave us a chance to pool our technical and investigative resources for a much larger aim: to support those who lack access to clean water. In particular, we wanted to build something that would empower journalists and civil society actors to press forward in this quest; as their storytelling and research into local water grievances can help activate communities for change. It was also important for us to test and leverage the expertise of the Code for All Network to overcome the data and resource barriers we were facing in our separate projects.

Building on previous projects

Though we never established clear cut roles, our organic collaborative style seems to be working thanks to the skill and commitment everyone brings to the table. There is OpenUp that developed Wazimap, a spatial open data portal to make geographically-linked information more readily available to non-technical audiences. A lot of rich demographic data and even some water-related information had already been collected into the platform’s central database, making it a snap for our tech team to fire up a new “instance” focused primarily on the new data we’ve been collecting for the Water Commons itself.

We greatly profited from the insights into journalistic needs and approaches provided by the Centre for Collaborative Investigative Journalism (CCIJ), a non-profit that was already producing several data-driven and visually rich investigations into water rights.

Code for Africa, an impact accelerator that supports a range of water focused data citizen science initiatives, has been working on water quality testing kits for water-borne diseases. Additionally, CfAfrica’s forensic data science team, the iLAB, helps transform this data into actionable insights for a network of watchdog CSO and media partners, performing the interface between both worlds.

And Code for Germany has dealt with water data on a more administrative level, trying to convince decision makers to open water data to the public. To demonstrate how this data can benefit the population overall, CfG developed a drinking water app that visualizes the quality of drinking water in municipalities throughout Germany.

First steps and problems encountered

We set up a weekly Jour Fixe (regular group meeting), which helped us gel as a team and feel comfortable discussing project issues as a collective. These standing meetings also helped us avoid most of the scheduling headaches that might otherwise plague a team of Africans, Europeans and Americans; though, as a kind of professional exchange bonus, we did learn about one another’s holidays, pets, snowstorms and even some construction work that occasionally made virtual attendance complicated. More beneficial, our weekly meetings have helped us minimize the duplication of efforts.

This was also part of our motivation when we asked Code for All members and others in our broader community to share any water-related projects they might have underway. Several colleagues pointed us to water data in their home countries and regions, and their responses in general underscored the importance of water to the civic tech community. However, existing project commitments and COVID-19 seemed to conspire against us as we tried to rope in more collaborators, especially when our work was just getting started and our ideas yet to take concrete form.

During these early days, we put a lot of effort into researching water data and hunting down any existing tools. No need to reinvent the wheel. But it became clear that we were in fact building something useful, especially if we could combine data from many countries and subnational levels in one place.

Image for post
Water samples taken in Makoko [Photo by Code for Africa]

Lessons learned

At this point it’s quite clear that there really isn’t enough water data available and that, yes, it is hard to find! Although international actors like WHO and UNICEF provide important global baseline information, subnational data is particularly scarce and rarely standardized. And water testing is still mostly a governmental domain, leaving few alternatives.

Simply put, gathering water quality data is an elaborate and expensive process. Because of the corrosive effects of water, you can’t leave toolkits submerged and unsupervised for long periods of time. Depending on the type of water body, samples have to be taken at different depths and locations and sample-takers have to be trained with the equipment.

With this problem in mind, we spent a significant amount of time searching for more basic assessments of water quality: photos for visible attributes, characterization of smell or even the presence of animals and microbes. We stumbled across the miniSASS project early on as a promising lead, which simply asks citizen scientists to collect samples of macro-invertebrates in any body of water to get a general measure of water quality. This low-tech, inexpensive approach piqued our imaginations, and when we suggested a potential partnership, the miniSASS developers showed mutual interest.

But even if we could measure water quality cheaply and uniformly, we uncovered other vexing problems. One of the biggest challenges facing journalists is simply a lack of localized data with the specificity they require. Water data is often presented at the aggregate level nationally, making it difficult to pinpoint problems within specific communities. And when we asked journalists in our network to describe the data they were after, they identified a dizzyingly diverse collection of needs: borehole locations in Nairobi; dam efficiency and water-level indicators; water pipe replacement and maintenance schedules; the cost, efficacy and byproducts of specific water treatment chemicals; and much more.

In Lagos, Nigeria, CfAfrica has been hard at work in the Makoko community tackling this information shortfall head on. The team trained groups of volunteers to take samples in local water sources and worked with local researchers to produce detailed lab results; all of which we are loading into our Water Commons database. The effort has given our team a good sense for how expensive and labor intensive water testing can be, but we’ve also seen an exciting level of engagement at the community level.

Image for post
Municipal Council meeting with the water volunteers
Image for post
The water volunteers with our samples [Photo by Code for Africa]

Meanwhile, CCIJ put out a call to its network of journalists asking them to uncover interesting datasets and pitch powerful accompanying stories. This approach ultimately forced us to think very broadly about what might constitute Water Commons data. Among the projects CCIJ decided to support: A photographic examination and data visualization of the environmental causes of land subsidence (literally sinking land) and its impact on one family in Gouda, Netherlands. In another story, CCIJ teamed up with a group of journalists in South Africa that has been scraping data from municipal websites in a bid to better understand the plight of the country’s overwhelmed wastewater infrastructure.

Working directly with journalists has generated a number of weighty questions. Do we include huge datasets, such as the satellite imagery being used to tell the land subsidence story in Gouda? What about other climate change data or information on health care institutions and labs? Would journalists want to measure a community’s access to drinking water by measuring, say, average distance to taps?

Journalists have also told us they want support gathering, cleaning, analyzing and visualizing data. How can we configure the Water Commons to meet those needs? We are now mulling over the need for complimentary online courseware that can offer assistance, along with the many visualization tools we hope to bake into the data repository itself.

What’s next?

Currently, we are in the midst of a massive effort to clean several of the datasets we’ve uncovered during our research and to upload them into a working prototype using the Wazi platform. Datasets must be transformed into collections of specific indicators that are tied to a geographic area; topics, country names and data formats are also appended. In this way, we are distinguishing between global data (more than three countries) and national data (fewer than three countries). Due to our broadened search, we have been able to cover a wide range of water rights’ indicators, from water quality measures, dam levels, water infrastructure assessments and more. But, as one might expect, we are locating data in a plethora of formats, from CSVs and APIs to PDFs and simple web interfaces that require scraping; and the data preparation work has proven to be quite time consuming.

Not only do we need to complete more of this work to make our prototype usable, but the team will need a good plan for how to incorporate more data down the line. And we need to tackle “front-end design” challenges going forward, including a better way to categorize and organize the data for Water Commons users.

Image for post
Water Commons logo
Image for post
Water Commons prototype interface

And we know there are other interesting ideas we’ll need to discuss; for example, we’re considering the value of creating and adding a virtual Rolodex, if you will, to the Water Commons to help journalists find credible sources and local water activists. Or, if we partner with the miniSASS creators, will we need to modify our work to better target groups like youth in general or Fridays for Future activists. In short, we’ll need to adapt the Water Commons to potentially meet the needs of a very broad audience.

While our Water Commons beta instance is already online, it’ll be clear to visitors that it’s truly a work in progress. With many of the challenges unresolved, ideally, we want to start another round of tests and collaboration with journalists, now with a more substantial tool already at place. As with most projects like this, more funding is needed. But if we get it right, we can build a tool — collaboratively and drip by drip — that helps fulfill the right to water.