My first time at a Code for Boston Hack Night was last December. It was right in the middle of the intensive data science program I was in. We had multiple labs due that week (as was the norm), and several of my classmates were taken aback by the thought of doing anything additional. But, having seen some of Code for Boston’s projects, I was itching to contribute and connect with others interested in using technology to address civic and social issues.
You can see additional info about the following projects on Code for Boston’s Current Projects page:
- Boston Info Voice App: An Alexa skill to answer questions about municipal services in Boston. Currently supports providing an address and asking for trash/recycling pick up days.
- Community Connect: A health resource web app that collects information about businesses and organizations that promote healthy lifestyle choices.
- MuckRock: File, track, and share public records requests.
- Safe Drinking Water Project: An exploratory data project to predict safe drinking water violations. It uses the EPA’s Safe Drinking Water Information System and the EnviroFacts API.
- Migrant Service Map: A web app designed to assist immigrant service providers by allowing them to create a profile, update information, and share it with clients quickly.
- Windfall Awareness project: A tool to help retirees affected by Social Security Windfall Elimination Program (WEP), a program which can reduce SSA benefits by up to 50% for certain public servants. This tool will help affected workers better plan retirement and self-advocate with the Social Security Administration.
- Plogalong: When you plog, you pick up trash as you go about your daily life… jogging, hiking, or simply walking down the street. Plogalong helps you track your plogs, connect with nearby ploggers, earn badges, and access local discounts.
It was a tough choice, but I ended up joining the Safe Drinking Water Project because they were specifically seeking people with data science skills. I also felt connected to the issue.
A story about a school
A few years ago, I was at one of the first meetings for a peer leader group at a local high school. While we waited for everyone to get settled in, I asked where to find the nearest water fountain. Several of the students raised their eyebrows.
“So I can fill my water bottle.”
“It’s down the hall. But you don’t want that.”
“Are you being serious?”
“Yeah, we don’t drink the water here.”
Everyone in the room nodded in agreement.
During the meeting, the students completed an activity where they identified issues in their school and communities that they were concerned about. As we wrapped up a lively discussion, I had to throw out one last question: “Nobody said water. Do any of you care about not being able to drink water in your own school?”
They all shrugged. It was such a norm that it wasn’t worth mentioning.
During the same week, Massachusetts Governor Charlie Baker announced $2 million of funding to help public schools test for lead and copper in drinking water.
Just over a year later, in May 2017, this article popped up in my feed:
The Summary Results of the tests confirmed that both of the high schools I was working with had “lead above action level.”
Understanding the data
This brings us back to the Safe Drinking Water Project at Code for Boston. The goal is to predict health-based violations in drinking water. We are starting with data from the EPA’s Safe Drinking Water Information System (SDWIS).
For me, understanding the data means more than knowing what information is contained in the SDWIS database (though we do finally have a centralized data dictionary!). It is also about understanding what that information actually tells us and identifying other factors that contribute to water safety that are not included in the database.
While I have been actively seeking out information to grow my understanding of issues surrounding safe drinking water, some of my learning has come from unexpected places.
AI in Flint
In early January, this article showed up in my inbox — not because of my tracking of water issues, but because of my interest in machine learning:
The TL;DR version of the article is that a team created a machine-learning model that was successfully predicting which homes in Flint were likely to have lead pipes. The City signed a contract with AECOM, a national firm, to speed up the work. Then:
AECOM discarded the machine-learning model’s predictions, which had guided excavations. And facing political pressure from some residents, [Flint Mayor Karen] Weaver demanded that the firm dig across the city’s wards and in every house on selected blocks, rather than picking out the homes likely to have lead because of age, property type, or other characteristics that could be correlated with the pipes.
After a multimillion-dollar investment in project management, thousands of people in Flint still have homes with lead pipes, when the previous program would likely have already found and replaced them.
The seas are rising
The week after that Flint article came out, I was at the Data for Black Lives II conference at MIT Media Lab.
The session that has kept me thinking about it the most since then was “The seas are rising but so are the people”: Data, Disaster & Collective Power.
The panelists — Valencia Gunder, Denice Ross, Lisa Rice, and Bina Venkataraman — shared their work and answered the questions: “What are the ways that Black communities are using social media, data, and technology to prepare for the next disaster? To prevent it? What are the role of data scientists and engineers in these rapid response moments and the movement for climate justice?”
They talked about the ways that climate change has exacerbated problems with higher housing costs and displacement. Valencia Gunder began her story:
As you do work in south Florida, climate change and resiliency is always part of the conversation, even in the most informal way. My grandfather told me as a child, “They’re going to come steal our communities because it doesn’t flood.” And I didn’t understand him until I was grown.
Others also talked about the connections between flooding and inequities in communities. The Data for Black Lives YouTube channel has a recording of the full panel, as well as other sessions from the conference, available to view.
Shortly after completing the data science program, I was talking with someone who made a tangential comment about a relative who uses artificial intelligence to look at how food for cows affects their manure, which in turn affects water systems due to runoff. He waved his hand in dismissal: “But you’re probably not interested in that.”
I explained that I actually was interested in the topic. This was a contributing factor that we hadn’t discussed yet in the Safe Water group. I did some digging to learn a little more about how manure and other farm runoff contaminates water.
A 2009 New York Times article considered the problem in Wisconsin, where “more than 30 percent of the wells in one town alone violated basic health standards”:
Runoff of waste from farm animals is said to be a source of pollutants in drinking water. At a town hall meeting, angry homeowners yelled at dairy owners, some of whom are perceived as among the most wealthy and powerful people in town.
It goes on to explain that even if local environmental agencies weren’t overtaxed, “a powerful farm lobby has blocked previous environmental efforts on Capital [sic] Hill. Even when state legislatures have acted, they have often encountered unexpected difficulties.”
Different political and economic interests are at play. And while Valencia Gunder talked about displacement in Florida due to flooding and climate change, maps like the one above (also using SDWIS data, along with data from the USDA’s Census of Agriculture) signal additional concerns. Increased flooding in areas already threatened by farm pollution raises the risk of contaminants in the water systems.
Toxic chemical sites
During Harvard’s Next in Data Visualization event earlier this month, Blacki Migliozzi, graphics editor at New York Times, shared some of his climate change visualizations. One, in particular, caught my attention: a map showing 2,500 chemical sites that are in flood-prone areas across the country.
This map was created using data from the EPA’s Toxics Release Inventory and FEMA’s National Flood Hazard Layer.
Similar to the issues raised by farm runoff, chemical sites in areas with more flooding from rising sea levels are also at higher risk of contaminating water systems, along with a host of other problems.
The Poisoned City
A couple of days later, another Safe Water volunteer and I went to see journalist Anna Clark speak about her book, The Poisoned City: Flint’s Water and the American Urban Tragedy. Granted, in this case, I went expecting to learn more about factors affecting drinking water, but it still feels worth including here.
After a brief intro explaining the city’s switch from Great Lakes water to Flint River water and subsequent issues with lead, E. coli, and Legionnaires’ disease, Clark asked, “What makes a city vulnerable to something like this in the first place?”
Her answer: “You have to go back a few generations.”
From there, her story started in the 1960s, “when Flint was the most segregated city in the North, and the third most segregated nationwide.” It included the impact of General Motors, the history of community organizing and the fight for fair housing, deindustrialization, and emergency management laws.
For an illustrated essay that covers some of this, see Clark’s collaboration with graphic artist Josh Kramer:
For me, two comments that Clark made during the conversation after her talk stood out:
Environmental issues always end in settlements with no admissions of guilt.
This wasn’t just a technical problem. There’s a reason this happened in Flint, and it wouldn’t have happened in Ann Arbor.
John, the other Safe Water volunteer who came to the talk, got a copy of The Poisoned City. Clark signed it to #water (our group’s Code for Boston Slack channel), and it’s now being read and passed through the group.
Meanwhile, this Twitter Moment was waiting in my inbox this morning:
This is yet another reminder of how differing interests, values, priorities, and allocation of resources can affect drinking water (and so many other things).
These were some of the connections I have made while trying to understand the data we are using. As always, I am happy to hear any thoughts on this.
If you have suggestions for other data, research, or approaches for the Safe Drinking Water Project to consider, please do share! You can also check out the Safe Drinking Water Project GitHub repo and join the #water channel on the Code for Boston Slack.
If you are in the Boston area and interested in getting involved with this or another Code for Boston project, learn more on the Code for Boston website and come on out to a Weekly Hack Night.