Stuart and Aurelia Moser ran a session about geocoders at Mozfest 2015, as part of the Science floor. Afterwards he connected with Nicholas from Where on Mars? and built a Mars Geocoder using USGS placenames for the planet, you can now get geojson for mars entities by completing the URL with proper names… Amazing and built in only a few hours at Mozfest.
Can you tell me a bit about your work?
I’m a researcher and a data scientist at Carto [now Head of Research & Data]. I specialize in mapping data online. We work with journalists and nonprofits. We take geospatial and turn it into beautiful maps and visualizations, and from that help uncover insights that can be used to make decisions.
What would be an example of a success for you?
Personally: making a difference. Helping make change in people’s lives and communities by helping them better understand their own data and open data available from the government and though other sources. Having an impact. Empowering others. My company also has this as a core value.
Why is open Internet important to you?
The open Internet is important because with geospatial data you have a lot of leverage. If you know where something happens you can tie in other pieces of data — or other datasets — by virtue of your location in a specific place. “Open” means others can join what you are doing and combine their data. This is important to providing context to your own data. For example, if you have survey data about people’s happiness or medical needs… when you tie that data to a specific place you can combine it with other measures, like income levels or diversity. This can reveal patterns. The open web and open data give you access to this broader context — for your own personal data and for others. Understanding context is valuable to you as well as to others.
How about an example of a challenge?
How do we make processes easy as possible for those who are not data scientists? We are working with complex algorithms, ideas, statistics, and geospatial processing. How do we take highly technical and academic process and make understandable and usable by people?
How have you approached solving this?
We try and remove as much decision-making as possible in the early stages. That means we have very good defaults — while giving people the ability to make changes to them later on. For example we took an academic geoeconometric method, Moran’s I, that is used to expose areas that are unusual. This is a complicated method. We spoke to researches who know it inside out to understand the defaults we might use as a starting point. Then we re-named the approach “Areas of Interest” and made the output more readable by assigning friendlier names. So we took something academic and wrapped it in language and code to make it more approachable. We brought it all together.
Another example. The US Census has a crazy dataset. Especially interesting is the US Community Survey, which asks a surprisingly broad range of questions that provide detailed insights into life in communities. The UK and Spain have similar. But the US Census dataset is huge — 150 gigs covering the entire US. But it is hard to read. The columns are numerical so you would need to look up what each column refers to in order to read it. You have to do a lot of work to use it. So we launched the Data Observatory. It lets people say what they are interested in, for example population, median income, number of people over 80. They can use this data to augment the information they already have. We made it easy to use. It is powerful and empowering. We’re making the data and functions more accessible.
But we don’t typically describe communities and neighborhoods using census variables. For example how would you describe Bushwick, in Brooklyn? So we created language-based descriptions by looking at all the variables and finding clusters. We found similarities and attached names to those. So far we have 55 categories to describe neighborhoods in the United States. That can be brought into a dataset.
I see you are helping others use data. It would be good to talk to some of your clients. To understand how this has helped them. Would that be possible?
Yes. Let me connect you with our grants team. We have a nonprofit grants program.
Can you tell me about how you got involved with Mozilla? What has that been like?
I’m on the periphery. Know people who work there. Meet them at events and conferences. Love what they do. I went to Mozfest in 2015. My last job was in Chicago, at the Zooniverse based at Adler Planetarium. We took scientific data that could not be processed by computers and built projects that invited people to come online and help us analyze data. We created 30 projects with 1.3 million contributors. This was science research. At the time I left, we has published 60 papers. One of my coworkers joined Hive Chicago. [A Mozilla Learning affiliate.] They funded some of our programs. Also I have connected to Mozilla Science in my previous career as and astrophysicist. I’ve been around a lot of people and done a lot of things with Mozilla without officially being part of the organization.
Can you tell me about a time that Mozilla had some sort of impact on your life or work?
Mozilla provided me with the opportunity to present about mapping techniques. I got to talk to others who are interested in this to pic. From those connections, we have started a planetary sciences project. We built a Mars Geocoder. You can ask about a place on Mars where that is. It will give you the US Government name.
We also partnered with Mozilla to run a science hack day in New York City. We combined this with a massive space apps challenge. Thirty people came and spent 40 hours working to enter a NASA competition. One of them even got to the semi-finals.
When I was working on my PhD in Scotland I would always talk to people about the universe, and this continued when I worked at Zooniverse and got involved in the citizen science world. Increasing participating in science is important. Helping people to think about and analyze data. This can help them make better decisions and improve their lives.
Can you tell me about a time that Mozilla did not meet your expectations?
Well, for the most part it has! Mozilla is good at creating networks and providing opportunities for people to meet and spark collaborations. The one example I can think of I think it was more our fault than Mozilla’s. We were not prepared so felt that the event, a Global Science sprint, did not have enough structure. But that’s actually what Mozilla does best. It helps people cross boundaries. At Mozfest we were in the science AND journalism tracks. And we interacted with folks from advocacy and privacy. It is rate that those communities come together.
Anything more you want to tell me? Anything you want to ask me?
It is impressive is how many people Mozilla touches who are not part of the core organization. I’m thinking about Hive Learning and Mozilla Science. Impressive. They are great at movement building. They help people hear each other’s voices. They enable cross-discipline collaborations. It is rare to have an organization whose focus is on generating those discussions and pushing those agendas. It is super valuable. Otherwise, it is hard for us to find time and breathing space to have those conversations and think about these things. Mozfest is important because it can advance these cross-boundary ideas. Mozilla creates coherent conversations.