EarthWise

Inspiration

Tackling sustainability and climate neutrality issues have been a passion of Sabīne for more than 10 years. Because of her working experience in the agriculture and forestry sector, it seemed like a logical target audience to develop some kind of solution. Especially because agriculture and forestry are industries that are highly affected by climate change and are also some of the main building blocks of our society. It is not only about economics but also about sustaining our primary sources of oxygen supply and biodiversity as well as feeding the World's population.

What it does

EarthWise is a land intelligence tool that helps farmers and foresters to manage risks and make sustainable data-driven decisions about their yearly and long-term actions in terms of irrigation/melioration needs, crop rotation and precise nutrient and pesticide doses. Our data service visualizes local changes in the area's hydrological regime based on data from IPCC climate scenarios and EUSPA Copernicus satellite program missions Sentinel-1 and Sentinel-2. This gives an insight into soil moisture, nutrient loss (and possible water body eutrophication), most appropriate crop species and needs for recalculating nutrient and pesticide doses.

How we built it

During the hackathon, we focused on researching Copernicus service data availability, usability and options for the data model and interactive map creation, not prototyping the end-user-friendly interface.

Challenges we ran into

Copernicus services Sentinel-2 and Sentinel-1 data free access is limited to downloadable files with raw data sets that must be wrangled before they can be used for data insights creation or ML model training according to our needs. Without large database storage spaces or cloud services, free access data sets can be utilized only for our small-scale product prototype creation.

The cloud services provide access to clean, sorted, real-time and historical datasets from Sentinel-2 and would be suitable for our needs including Planet Agricultural and land datasets and Sentinel-Hub datasets. These cloud services require subscription-based payments to sustain the infrastructure of the data storage and frequent calls via API.

There is still a need for more research on what additional data should we use to provide a user-friendly interface and actionable insights from the data points of our interactive map e.g. the option to overlay data from the address-based map like Google maps for address lookup.

Accomplishments that we're proud of

Sabīne as a team lead and domain expert is super happy that she found Ieva who is an enthusiastic, passionate and very professional data scientist. We are proud of our match and that we defined an exact value proposition, and developed a technical scheme for our solution.

What we learned

By utilizing the methodology and mapping approach behind IPCC report creation based on publicly available GitHub repositories we can build our own interactive, visualization map of the boreal biogeographic region - Baltics and Scandinavia - with our predictive modelling, acquired data sets and calculations that would help farmers to make data-driven decisions for sustainable production.

The use of Sentinel-2 and Sentinel-1 data will be carried out in two phases:

Phase-1 Create a small-scale prototype with a working interactive map that would combine features like - searching for the farmer-specific region by address, prediction of how the soil structure will change in time depending on climate change scenarios and actionable advice on how to farm effectively and sustainably. The prototype would be based on available climate change scenarios, freely available downloadable data sets from Copernicus service and other region-specific data with an adapted approach from IPCC report visualisations and data mapping.

Phase-2 Based on the working prototype example create the full-scale interactive application with the same functionality using Sentinel Hub cloud services provided API gateway and Cloud-based ML services for predictive data modelling creation. The solution based on cloud service providers would be the beta version of the end tool and would be tested by farmers who agree to become our beta testers.

What's next for EarthWise

Since EarthWise as a concept is live for about 4 months now, the next steps are introducing Ieva to the whole EarthWise team (planned on Monday, 13th of March), validating the solution in client interviews, applying for a grant to develop our first functional MVP, building and iterating it, and looking for seed investment to launch our product.