Seven things we learned at Geo for Good

By Raleigh Seamster, Program Manager, Google Earth Outreach

Earlier this month, Google Earth Outreach hosted 200 mappers from across the globe at our California headquarters for 4 days of hands-on trainings in our mapping tools. The attendees came from diverse backgrounds — nonprofits, scientists, educators — but they had at least one thing in common: they all leverage mapping tools for social and environmental good. Their work is the inspiration behind the name of this annual event, now in its 7th year: the Geo for Good User Summit.

1. The Geo for Good community can influence the direction of Geo tools

As we filed into the main room and found our seats, our director Rebecca Moore took the stage to welcome 200 members of the public benefit community to Google. Rebecca highlighted different Geo tools that attendees would be learning about during an intense week of training sessions. She talked about how our user community had influenced Google’s motivation to build Google Earth Engine: While we were visiting the Amazon. Brazilian scientists explained that we’re losing a million acres of rainforest every year, and a lot is happening in remote areas where there is no law enforcement on the ground. But, with daily satellite images coming in and running change detection algorithms the scientists saw there was potential to build a virtual deforestation monitoring and alerting system. But the problem was such a system would require petabytes of data and no normal institution could support such a project. The scientists said — build something that lets us do science and analysis at scale. That was the origin of Earth Engine.

Rebecca reminded Geo for Good users that they’d be learning about all the different Geo tools this week and they should let us know what works well and what needs improvement:

We build these tools guided by input from the community….don’t be shy about talking to us because that’s how the tools get better.
Participant giving feedback to the Earth Engine team during Office Hours

2. With BigQuery, you can now ask even bigger (geospatial) questions about big data

One tool that was taught in one of the several breakout sessions was Google BigQuery’s new geospatial features. Google BigQuery is a tool within Google Cloud Platform which can handle big data — petabytes of data — structured in a tabular format. It’s great for storage, querying, and analytics of big data, and over this past summer, the Cloud team announced that BigQuery could now store vector features — points, lines, and polygons — and allow you to run large GIS analysis over the data, such as intersections, unions, and other spatial joins and queries. The participants queried weather and air quality monitoring stations, finding observations closest to a given zip code, and then visualized their “big queries” with the BigQuery GeoViz tool.

Participants used BigQuery Geo Viz to map results from their BigQuery GIS query. This map shows air quality monitoring stations that are within 50km of the Summit location.

3. There’s a new tool that addresses the challenges of mobile data collection from a map-first perspective

In another breakout session, Summit attendees were invited to play with a prototype of a new open source mobile data collection tool called Ground. We had a mission: find rare species of frogs in the wild (i.e. look for toy frogs cleverly hidden under bushes and in tree branches around the Google campus), add its GPS location, and answer a few questions about the frog (toxic or not?). As we all started finding frogs, we saw the map on our Ground app start to fill up with frog placemarks, which were later displayed together in Google Earth! At the end of the session, we got to have a voice in the future development of Ground as we voted for must-have features. People are really looking forward to the release of this new tool!

Ground participants voting on the features that were most important

4. National Geographic Society and Google are working together to achieve a Planet in Balance.

Returning to the main stage, Dr. Jonathan Baillie, Executive Vice President and Chief Scientist at National Geographic Society, talked about the challenges of measuring how our planet is changing through time. For example, we’ve mapped where the world’s protected areas are, but we don’t know if they actually working. “We’re in the dark ages of biodiversity data and we’re just slowly crawling out,” he said. He talked about how change is coming about due to a massive revolution in technology: more wildlife tracking data than we ever had before, major breakthroughs in sensors, citizen scientists contributing biodiversity data from their mobile phones, technology enabling predictions of the health of our planet.

Then, Dr. Baillie announced a partnership with Google to monitor vital signs of the planet, understand threats to the natural world and realize solutions to help achieve a planet in balance. Through the partnership, Dr. Baillie hopes to inspire action by educating consumers and decision-makers about the critical importance of protecting at least 30 percent of the planet by 2030.

5. Earth Engine ❤ TensorFlow ❤ Sentinel-2

Stepping into the breakout room Coast Live Oak, we found small groups huddled around desks — it’s the Hackathon! Sitting down with one of the groups, we learned why they were hoping to locate small dams. Charlotte Weil of the Natural Capital Project at Stanford University explained to us that “millions of dams and reservoirs have been constructed worldwide to manage water, however they can have big environmental impacts on freshwater habitats and fish migration. Knowing the location of these dams and reservoirs is crucial to protecting these ecosystems however we only know where the largest dams are — the smaller dams are usually not mapped.” This Hackathon team took on the challenge of building a deep learning image classification model to locate small dams and reservoirs using Earth Engine, Sentinel-2 and TensorFlow. For a deep dive into the project, check out Charlotte’s post.

Hackathon team using Earth Engine, Sentinel 2 and Tensorflow to find small river dams

6. “Together we can finish the map”

Back on the main stage, Street View Product Manager Stafford Marquardt talked about the history of Google Street View. Street View has been around for 11 years, and during that time it has grown from a project to map a few cities to imagery that covers over 100 countries — but there are still many places Street View hasn’t mapped. Stafford noted that audience members are a diverse group of people who are passionate about protecting the natural resources and cultural heritage of some of the most remote and inspiring places on this planet, some of which Street View has not yet reached. That’s why we’ve opened Street View as a platform where anyone can contribute, so we can build this virtual library of the entire planet together.

Following Stafford on the stage were three organizations who talked about how they’ve used Street View to map unsung places from around the world, including national parks in Canada, coral reefs and chimpanzee habitats in Tanzania. Check out the full recording on YouTube.

Jeff Bolingbroke of Parks Canada talks about how they’ve mapped 3,381 kms of National Parks, Historic sites and Marine Conservation Areas with Street View

7. Tools like Google Earth can be used to put underrepresented communities on the map

Claudinete Colé and Rogério de Oliveira Pereira traveled all the way from the Brazilian Amazon to the main stage at Geo for Good to represent the Association of Remaining Quilombo Communities of the Municipality of Oriximiná and their mission to put their people on the map for the very first time using Google Earth. Quilombolas are the descendants of African enslaved peoples who escaped and formed independent settlements. Today about 1.5 million Quilombolas live in approximately 5,000 communities in Brazil. Only about 10% of these communities have earned legal recognition of their traditional lands and Claudinete and Rogério are working to empower Quilombola communities to map their lands for the first time and use the data to better manage their lands and gain land tenure.