Guardian analytics: labels and top down data way points

[anvabr]

Problem description

Guardian policy flows enable project developers, VVBs and other participants to effortlessly follow the correct methodology when reporting and validating environmental projects. Although they generate a significant amount of verifiable project data, auditors often require further assessment of these data in the context of various external context/information, which may or may not be machine readable and/or easily accessible to the public. Furthermore, the assessment procedures may themselves be formalised into complex frameworks combining quantitative and qualitative methods. One example of such is Core Carbon Principles (CCP) methodology. At present Guardian does not offer functionality to enable such workflows.

Note

#3341 comments section contains the original discussion on this topic. The scope of that original ticket is broad. This ticket contains specific immediate scope of 'top-down' analytics to enable the support for CCP and similar methodologies.

Requirements

• Introduce 2 new workflows into Guardian, which include the corresponding roles and access permissions:
  • labels author, for users to be able to create the 'rulesets' for evaluating data for their compliance with the chosen 'label',
  • auditor workflow, for users which would use these 'rulesets' to apply to data.
• Introduce the concept of labels, which can be specified to combine multiple statistics (introduced in Guardian analytics: bottom-up data traceability #3336) to create 'higher-order' statistics which themselves can be combined further essentially enabling the creation of 'data transformation' trees which, when applied to data, would ultimately get resolved into binary compliant/non-compliant answers. The top-level 'nodes' in these trees are 'Labels'.
• Enhance the current capability of qualitative evaluations in Statistics to support the ability for users to attach external evidence and add textual comments/explanations whenever a human input is enabled. The evidence would then become part of the 'evaluation trust-chain', i.e. it should be hashed and stored verifiably. Evidence in the image formats should be viewable in the browser, archives (zip files), pdfs, csv files should be supported for attachment and then download.
• Enable Auditors to apply 'label rulesets' to tokens, Guardian would then automatically traverse the token trust-chain to find and evaluate the required data to produce the label conclusion, i.e. the compliant/non-compliant results. These results can optionally be published to IPFS/topics by Auditors that generated them.
• Enable ordinary users to search for statistics, label ruleset, and label conclusions that have been published.

Definition of done

• Labels functionality is operating as described above
• Documentation and examples are provided

Acceptance criteria

• An example/test case for CCP evaluation using Labels functionality can be demonstrated