Soil ID Algorithm

[ltseng]

The core of the value add provided by LandPKS. The Soil ID algorithm provides the user with predictions about the type of soil at a given location, and provides increasingly better results the more information is provided by the user. The SoilID algorithm is deployed in multiple user stories, including Soil ID tiles, Soil ID dashboard, and Soil Info screens for Sites and Locations.

References

Product Spec: Soil and Ecological Site ID
Figma
Soil and Ecological Site ID
Paul's implementation notes

Backend

Beta Give feedback

1. Create SoilID API end point - Dummy data #617
   0 of 2
   Engineering +1
   
2. Load and display Soil ID data in mobile client location screens #653
   Engineering +1
   
3. feat: define data schema for non-site soil match GraphQL query terraso-backend#1265
   Engineering +1
   
4. feat: define data schema for site soil matches GraphQL query terraso-backend#1266
   Engineering +1
   
5. discovery: attempt to integrate soil ID algorithm code into backend codebase terraso-backend#1267
   Engineering +1
   
6. feat: hook up location based soil matches graphql query to real soil ID algorithm terraso-backend#1269
   Engineering +1
   
7. feat: hook up data based soil matches query to real soil ID algorithm terraso-backend#1270
   Engineering +1
   
8. Placeholder - soilID algorithm work terraso-mobile-client#1430
   Engineering +1
   
9. test: create snapshot tests to validate the output of the soil ID algorithm terraso-backend#1268
   Engineering +1
10. Use Soil ID algorithm for temporary site popover terraso-mobile-client#1555
    Engineering +1
    

Acceptance Criteria

• Soils related components will be able to access the following information
• Location SoilID
  • Accept: Lat, Long
  • Return: Top 12 matched soils: Name, Percentage Match
• Site SoilID
  • Accept: Lat, Long, multiple depths of texture, rock fragmentation, color, slope
  • Return:
    • Top 12 matched soils: Name, Percentage Match
    • Recommended next inputs
    • Functionally equivalent soils
• all queries will take <5s