Product requirements

Goals

Farmstack should help organizations and individuals across the agriculture value chain to share data and realize more value, more here

Strategic fit

  • DG is established partner in various state government’s agriculture programs in digital learning space involving video dissemination/ digital advisory

  • Data sharing infrastructure augments and leverages on the existing efforts in digital advisory and fits with the mission 

Assumptions

  • Orgs and people don’t trust DG or any other org to host data

    • validated in discussion with different people and also reports of farmers mistrust

  • Orgs see value in data sharing for knowing more about farmers

    • signals that relevant farmer profile not available with anyone and it helps build better solution

  • Current initiatives fail to provide necessary trust across ecosystem

    • signals of mistrust with the initiatives that tend to be centrally orchestrated

Customer Development Process

User concerns - User requirement Matrix 

User concerns

Possible answer - user requirement

Priority rating (1-5)

  • Excited

  • Need

  • Intensity (# orgs)

Total: (E+N)*I

Trust

  • I share data but don’t control usage

  • How do I enforce legal contract

  • Policy based usage control

 

  • 5

  • 1

  • 4

T: 24

Data security while handling data

  • Who hosts the data?

  • Why can’t it be me?

  • Do you see my data?

  • Peer to peer data exchange

  • Data should be exchanged from org A to B directly

 

  • 4

  • 3

  • 3

T: 21

Incentive

  • Why do I share using farmstack?

  • p2p with usage policy gives you complete control

  • prepare use cases to show value

  • data catalog helps discover new opportunities

  • 4

  • 3

  • 3

T: 21

Data Policy

  • Some concrete like GDPR

  • Some not concrete

  • Mention of FAIR data principles

  • p2p connector with usage policy gives compliance to GDPR by design

  • data catalog also makes data FAIR

  • 3

  • 4

  • 3

T: 21

Search data

  • Can I search descriptions of data before initiating data transfer?

  • Data description catalog

  • 4

  • 3

  • 2

T:14

 

Quality assurance

  • how do you say that the data is correct and collected properly?

  • who was the eventual owner?

  • any reports or certificates

  • Data quality report/certificate is another data exchange, can be linked to actual data

  • farmstack can aid provenance, not provide it

  • 3

  • 4

  • 2

T: 14

Open Data but lot of effort spent in pre-processing data

  • Discover data services

  • Trusted source to download and deploy existing data transformation services

  • 2

  • 3

  • 1

T: 5

 

Further Questions

  1. Who owns FarmStack (p2p data exchange tool)?

    1. FarmStack is operated by Stewards and participants based on the open source repository

    2. DG owns FarmStack open source repository with community of participating members

  2. What policy based usage controls are required immediately (6 months)?

    1. Ability to prevent data to be downloaded

    2. Ability to expire data

    3. Use data for processing but not available for further use

    4. Ensure that data is being processed/consumed within the geographic limits

  3. What level of user friendliness is required in defining usage policy?

    1. Suite of policies that can be selected as well as option to easily define usage policy for specific cases

  4. What is the level of security?

    1. Entry point should be security by obscurity - have virtual environments

    2. Depending on the criticality of use cases, security can be addressed

  5. Who are the power users?

    1. Those who are willing to run the network

    2. Those who are willing to manage the connectors

Overall Plan

  • Track 1: Tech development focussed on priority a) p2p connectors, b) policy based usage control and c) data discovery

  • Track 2: Use case development with customers (focus on diversity in value chain, data patterns, applications that help farmers in some way)

  • Track 3: RnD and collaboration related to backend

Not Doing

  • Security centric design: Stick to security by obscurity as starting point and then see what else is required

  • Application heavy use case: eg, a novel application that involves AI

  • P2P Ag Data marketplace: FarmStack is not a marketplace

  • Improving data: FarmStack is not a tool for improving quality of data