Create Metabase dashboard for documenting assumed invariants in Littlepay data

[mjumbewu]

Is your feature request related to a problem? Please describe. Littlepay data has several apparent integrity issues.

Describe the solution you'd like We can have a dashboard where we can lay out the things we assume to be invariants in the Littlepay data. If there are things that violate those invariants, we know we have a problem. Likewise, if there are no violations, then we can assume that our invariants and the data are in line with each other.

Describe alternatives you've considered We already have a notebook to run tests against Littlepay data, but a dashboard is an easier way to get a quick view on our assumptions.

Additional context Assumptions/metrics:

• We assume that device_transactions.littlepay_device_transaction is a unique ID. How many littlepay_device_transaction values are not unique?
• We assume that the micropayments.micropayment_id field is a unique ID. How many micropayment_id values are not unique?
• We assume we can disregard route_id = 'Route Z'. How many device transactions only have Route Z data?
• We assume customers should be identical between micropayments and device transactions. how many micropayment/device transaction pairs have mismatched customer_id values?
• We assume that multiple device_transactions associated with a single micropayments record implies a tap on/off situation, and that these trips should have device transactions labeled as tap on and tap off, How many apparently variable-length trips have device transactions labeled as single?

Filters:

• Provider ID
• Exclude "Route Z"

[mjumbewu]

We now have a dashboard to track data validation issues. Most of the charts on this dashboard should be empty, with the exception of the Route Z question.

https://metabase.k8s.calitp.jarv.us/dashboard/28-payments-data-validation