Learning from Earth Index users

mikel
Earth Genome
Published in
4 min readNov 19, 2024

We surveyed Earth Index users and our top learnings are:

  • Don’t shy away from more technical features, while still building with maximum accessibility in mind
  • Assume some level of “geographic understanding” in our users
  • Food systems and nature based solutions use cases are most ripe for deep dive on additional market intelligence.

As we build Earth Index, we are staying very closely tuned into what our users’ needs and expectations are for Earth Observation and AI tools, and what use cases and applications Earth Index is in the best position to address. Last month we began moving interested people off our waitlist and granted access to Earth Index, and we are so excited to see what they find, and learn what’s working for them and what more they need.

How did we survey? We took the opportunity to dive deeper with waitlist participants, who had already shared some details on what they’re searching for. With the expert guidance of Deepti Illa and Sridhar Guthula of Thedas Ventures, we designed and conducted a user survey following the alpha release of Earth Index earlier this year. Our mission was clear: uncover the most compelling opportunities for Earth Index, identify pain points and concerns, refine which use cases resonate (and which don’t), and understand how Earth Index fits into the broader ecosystem of tools, data sources, and integrations. These survey questions were mapped to different points in the “customer journey”, to see the bigger picture of their thinking, actions, and goals. We’re deeply grateful to the 40 individuals who generously spent their time answering our 20 questions. A special thank you goes to Deepti and Sridhar for their expert analysis, which brought invaluable clarity and direction to our findings.

Who responded? Respondents came from every corner of the globe, and a full range of types of organizations. What was consistent was an existing technical “bent”, with experiences in data science, machine learning, and environmental science, and profiles like scientific researchers, geographers and software developers. Typically these folks are already using cloud based and desktop geographic data tools, leverage programming languages and other frameworks for data analysis, and seek and have access to many sources of environmental data. While not everyone was deeply technical, everyone came with an understanding of what spatial analysis and data could do for environmental problems they are trying to solve. There is certainly some sample bias among these early adopters who signed up for the waitlist and responded to the survey, but we believe the insights gained remain highly valuable. We will build Earth Index for Earth monitoring for everyone, but can assume users already have some practice of spatial thinking, and that even the most tech savvy appreciate getting a leg up and avoiding the hard work of organizing global satellite imagery and AI models.

What are they trying to do? Users are looking to develop tailored solutions for sectors such as agriculture, urban planning, environmental monitoring, and disaster response; all objectives aligned with a wide range of familiar environmental challenges. Monitoring restoration and conservation projects, including as input for financial instruments. Crop detection and identification of other points on the food supply chain at processing and distribution sites. Detecting illegal mines. Users also want to monitor climate resilience (identify areas at risk, develop strategies, evaluate the effectiveness of recovery efforts , predictive modeling to know how well the environment will withstand adverse conditions or disruptions). Tracking the flow of water and moisture on the landscape, and water management infrastructure. Monitoring change offshore and in maritime environments. Even some use cases around education, teaching about Earth observation, AI and the environment. One surprise was detecting volcanic activity and geological change!

What are general pain points with AI and EO tools? One of the most common barriers to applying AI to Earth Observation is the high compute and maintenance cost of managing the data pipeline to download, process, model and index with imagery. These are simply hard to access large data sets, especially when high resolution. And it’s not a one time process — to be responsive to a changing environment requires recent imagery. Frequently this is all made more difficult by low connectivity environments. And even then, it can be difficult to interpret satellite imagery, accuracy of automated detections is a concern, and there’s a high technical learning curve to apply machine learning.

What features are important? When looking at what Earth Index offers, a number of features stood out as important to cover. Global coverage of indexing was a top need (and is requested regularly all around). Ongoing monitoring and change detection is just as high a need. Availability of pre-seeded shareable searches would make it easier to get started with on a new search. Related, guides and examples of other specific applications of Earth Index would help for learning to build successful searches. Integrated solutions, that fits search results into a bigger data workflow and product development, would help make usage immediately impactful. For more technical users, there’s a desire for direct API access to build searches into other experiences, and support for advanced model training directly in the UI.

What concerns do they have about the Earth Index? Is Earth Index free or how much will it cost to use? How reliable is the data quality? How easy is it to use? How mature and stable is the product? Will there be ongoing support for using Earth Index into the future?

A lot for us to reflect on! Both great confirmation of the direction we’re heading and the challenges we are eager to work on. More soon on our plans for the year ahead. Are you interested in the Earth Index? Sign up on the waitlist and we’ll grant access soon, as we’re increasing the pace of releasing people off the waitlist. And get in touch to discuss what you’re working on and how Earth Index and Earth Genome could help.

--

--

Earth Genome
Earth Genome

Published in Earth Genome

We build technology to create data, apply science, influence, guide and measure effective actions on the most pressing climate and conservation issues of our times.

mikel
mikel

Written by mikel

Mapper. Coder. Earth Genome. OpenStreetMap Foundation. HOT. former Mapbox / Presidential Innovation Fellow. Views are my own.

No responses yet