Tech Roadmap

Organized around key deliverables:

Collection and Verification: How do we make it easy to record what practices farmers are following and validate that what is reported is accurate? Initially start with VRPs reporting with JEEViKA + DG team validating and move to a model where farmers self-report and VRPs validate
Storage: What is the underlying data model for all the practice data that is captured? Want to enable the same underlying data to be re-used across use cases.
Model Integration and Output Visualization: How do we feed the various climate models to generate emission estimates? How do we present model outputs to enable inter-comparisons?
Data Wallet / Data Sharing: How can farmers grant informed consent for their data to be shared with external stakeholders?

	Current State (July 2021)	Future Plans

	Current State (July 2021)	Future Plans
Collection + Verification	VRPs record practice data via a mobile (built on Kobo/ODK). Form fields have been developed based on input requirements for climate models. Skilled Extension workers (SEWs) and Master Resource Persons (MRPs) along with DG team validate by calling and visiting farmers See data verification protocol for detail	Evolve to a model where farmers self-report and VRPs validate through calls and field visits Leverage Kisan Diary for farmers self reporting practices Align VRP payments and performance management to accurate data collection Integrate third party datasets (eg, remote sensing) to reduce cost of validation while preserving accuracy
Storage	Data captured via form is in a single table. See sample data file.	Evaluate FarmOS as a back-end store
Integration and Visualization	Re-structure data collected from the field so it can drive three models: DNDC, CoolFarmTool, CENTURY Publish model results in a dashboard	Develop automated data pipelines for practice data collected in the field to the various climate models; look at complementary OpenTEAM efforts on this (FarmOS → COMET/ DNDC/ CFT)
Data Wallet	Research to understand who will be consuming the data and user requirements around informed consent Reference this writeup for further detail	…

MRV