Creating Technology for Social Change

Data For Equity: The Power of Data to Promote Justice – Liveblog

This is a live blog account from the Data For Equity: The Power of Data to Promote Justice event.


Barbara Best, Executive Director, Center for Public Leadership, introduces the panel. The moderator is Yeshimabeit Milner, the Executive Director and Founder of Data for Black Lives which uses data science to create concrete measureable change in black lives.  The panel has been organized by Black Student Union and Black Policy Conference. People are tuning in via the livestream.

Yeshi introduces Data for Black Lives. They are building a movement for scientists, activists and organizers. We can use data and tech to make concrete change in the lives of black people and all people. Data and tech is changing the world so fast. We can look to the past to respond to the present moment. In 1793, Eli Whitney invented the cotton gin – separated seeds from cotton fiber. Cotton became king in US. By the 1850s, the Us produced the vast majority of cotton produced worldwide. The cotton gin was a gamechanging social invention. But it had extraordinary negative impact on transatlantic slave trade. For millions of enslaved people, cotton gin helped expand a cruel and violent system. No technology is neutral. For far too long data and tech have been weaponized against community. But we have examples of technology for positive social change. We are seeing advances now in civic analytics and data at all levels of government. Data plays a huge role in allocation of resources. These tools have a role to play in equity and to help elevate the voices that have been silenced.


This call to action is more urgent than ever before. How do we use data to expose inequity and hold governments accountable?


The four panelists here have used data in inspiring and different ways to promote justice. They include:

William Isaac, Fellow, Open Society Foundation; Research Advisor, the Human Rights Data Analysis Group
Kelly Jin, Director, Data-Driven Justice at the Arnold Foundation; former Citywide Analytics Manager, City of Boston
Carlos Rojas, Special Projects Consultant, Youth on Board; Founding Member, Boston Education Justice Alliance
Paola Villarreal, Harvard University Fellow, Berkman Klein Center for Internet & Society; former Data Science Fellow, ACLU

Yeshi will use some of the questions submitted online prior to the event to guide the panel discussion. Each panelist will briefly introduce themselves.


William Isaac leads with an introduction of his work. He focuses on algorithms and their role in public decision making. There’s an assumption amongst policy-makers that data is good and objective. Through his research he wants to show that data is not objective and that algorithms in and of themselves do not solve those problems. His research tries to illuminate that and build towards something better.


Kelly Jin has worked in many organizations. She feels everyone should tackle one issue: every year we have millions of people cycling through local jails. This costs a lot of money. When you look closer many have mental health issues, substance abuse issues. They launched the data-driven justice program at the White House last year to try to address this at the local level. People keep ending up in jail. How to bring together ER doctors, sheriffs and local communities to hold government accountable? She previously built the data team at City Hall Boston.

Paola Villareal says she comes from a corrupt country and she thought it would be different here. But it happens that it’s not – it’s just a different kind of corruption called oppression. There are many pipelines that show that. Every state and city has a different type of oppressive pipeline that is biased against people of color. She is here because she started to work on data, analyze it and show disparities. It is super important to learn about these shocking biases and oppressive systems.

Carlos Rojas says he comes from a perceived corrupt country – Colombia – and moved here when he was five. They moved to Dudley Square. After 6 months here, he became undocumented because they had flown in with a tourist visa. He noted how black and brown kids interacted with police. He was told to never, ever approach or talk to a police officer. If you do, be very polite or you might get arrested. As he grew up, he became aware of the ways that these problems start within the school system. What does it look like to reform school-level policies. School-to-prison pipeline in this country is a real thing. It sends young black and brown people from school directly into prison. They believe that young people organizing and in partnerships with adults can make beautiful things happen. Data that corroborates that lived experience is also very important. He has examples of amazing advocacy efforts but they are also having struggles getting the state to hold agencies accountable.


Yeshi says this represents real breadth. One of the first questions from the audience – what have been the pro and con impacts of data-driven decision making in government over the past decade?


Kelly says she wouldn’t go that far back. The core of a of the work around open data has only blossomed in the last five or so years. Cities have done a lot of work to open up their data. It’s hard to unlock data, but fundamentally this is public data. That’s the first step – how do we unlock it. From that, the engagement of a much broader community is what matters. If you have 300 data sets it means nothing if no one is using them for policy change or research or recommendations. The policy changes that have happened as a result of people looking at the data are what matters. One question is what tech infrastructure can we build on top of open data to provide value back to citizens. For example, individual health data donated to researchers. Technology – why aren’t we using more open source? It’s amazing to see the turnout for this event – how can these people get involved?


Paola says that people are here because of openness – this has happened in the last 8-10 years. Open source, open government, open data. It’s not just a set of tools but a mission. Openness in general is one of the best things that have happened. But on the other hand, machine learning in the criminal justice system is one of the worst things that has happened. We need to solve that. In the meantime, I would call for an embargo on that.


Walter agrees that openness is the biggest thing he has seen. We have seen the big coastal cities who have embraced data. He has seen something different in the midwest. In Michigan, they faced the Flint water crisis. They had no digital records of the water pipe existing. It turns out that they actually did have records but they were on file cards and there was one person responsible for them. Loveland Technologies is a company that then took those cards and digitized them. Data does have a lot of good use when you are allowed to share, usually with partners not inside government. The cons of this movement – there have been some weird things coming out of machine learning.  Part of it is algorithm but part of it is the institutional decision making on top of that. Some people in government don’t want to use data at all. Others say data will solve all our problems and don’t like when you say bad things about data. There has to be some middle ground. It’s not just machine learning or algorithms. Some places have predictive policing but don’t even use the software. Even when they have the tools when the institutions don’t change as well then nothing changes. Particularly need to focus on accountability mechanisms.


Kelly adds that one thing they have talked about is TQ – what is the technology/data quotient within government and how do you improve that? Data and tech vendors come into government. What are you doing? What decisions making? Data ethics and algorithmic transparency – how do you ensure that? The algorithms should also possibly be public.


Carlos says that in the school systems it has been striking what data has been capable of both positively and negatively. No Child Left Behind put schools in frenzy of data collection around test scores and what amounted to a toxic culture of high stakes testing. The policy ended up doing exactly what it was not designed to do – closed schools, left many students behind. The biggest predictor of how well you do on a test is how much money your parents make. But on the other hand, youth organizers have been demanding that schools collect more data. In the case of dismantling the school to prison pipeline, the state wasn’t collecting data on school discipline. We didn’t see data on who was being suspended and expelled. For years, we had to rely on personal narrative and anecdotes to prove that young people of color were being suspended at egregiously disproportionate levels. We demanded that the state collect disaggregated data, school by school, so that we could see better the school to prison pipeline.


Yeshi asks the next question. Openness has made it possible for us to be here today. But one thing we are grappling with, once people get the data, what are they going to do with it? Not everyone can learn R. She got involved with data collection as a youth organizer. How do we scale data literacy and change how we teach about data to get more women and people of color to make the open data movement more lively and accessible to more people?

Paola says we need more ways to tell stories and show how data impacted communities. We need more community engagement and co-creation. Show communities the data and ask them what they think. Data scientists are not saviors. They collaborate with communities to define the problem.

Carlos states that he has seen the impact when data scientists align with community groups. And they have also seen data waste when researchers create data and then it just sits on a shelf and isn’t used by groups that could benefit from it. They are lucky to work in a city that is rich in data science. Lot of people in Boston are interested to take direction from people on the ground. They come to them and say “What research do you need done to make your work effective?” Then when young people are in a legislator’s office they have data to back their arguments. We have been very invested in those partnerships. On Oct 19th, they will be gathering with youth and parent groups to review at the Chapter 222 data, talk about what is happening on the ground, determine how to move forward.

Paola says the most impactful work she has done was in teams of lawyers, community members, advocates and activists.


Kelly says she wants to talk about the role of media in this. How do we continue to show that there are women and minorities working in this field? For example, the film Hidden Figures. And on “how do we engage” – not to have data for data’s sake but to determine what the questions are and then figure out what are the data sets to help make those easier to answer. What data are we not collecting? There are so many cases where no one is collecting that information. There is a huge piece of catching up.


Walter thinks a lot about how you teach these concepts. For undergrads and people in college they created InnovateGov program that teaches them data science and then places them in government agencies. They found that you need to have something that you are passionate about. A lot of the stuff is boring. But the coolest part is that when you possess the knowledge you can present it to someone to make a case that they should change things. For example, a student team collected surveys about how to reach people involved in foreclosures. For high school students – smaller toy data sets where you introduce concepts and giving them passion or interest in a topic.

Yeshi asks what are some good examples of cases where agencies and orgs have used data for justice? That will help us after the Q&A.

Carlos talks about Youth on Board. They created surveys with questions that would help paint picture of students of color as well as listening projects where they would go have quick conversations. To engage them, they needed two things: big signs and bags of candy. Their questions were like, “Do you have police in your school? How do they affect the environment? Have you been suspended? Do you think it was fair? Do you think your race had anything to do with it?” We didn’t find anything surprising. They passed the Chapter 222 legislation which said instead of districts just doing zero-tolerance, they had to try other methods before suspension and expulsion. But they had no way of holding schools accountable on this. They decided to develop an app that summarizes major changes and allows you to report an equity grievance and they developed the Boston Student Rights app. Incredible tool that collected 26,000 cases. Students are using it to educate themselves and their teachers. Sometimes they are using it to advocate for themselves and prevent themselves from being suspended. All the data goes to the department of equity which the community group has a good relation with.

However, but now the schools are doing things like dismissing students early, doing emergency removals, and providing an informal no-trespassing notice. These things fall under the radar but then are not held accountable from the state.

Paola discusses her work for the ACLU and its relationship with the City of Boston. Although the entities did not agree, there was an open and transparent process. Data & Society is a great research organization.

Kelly talks about Measures for Justice – they are doing the hard work to do data collection and make it open and available. How do philanthropists step up to do that work? Coding it forward class at HKS taught by Nick Sinai creates partnerships between undergrads and the city of Boston. Finally, Jen Palka runs Code for America, like Teach for America for technologists who are placed in local jurisdictions.

Walter says he has so many examples. Sam Singyawe is amazing and part of Black Lives Matter started Mapping Police Violence. Some smaller ones in Detroit – like Data-driven Detroit, Future City Detroit – projects that are building the public infrastructure for data. A lot of nonprofits are doing the heavy lifting.