As a PM for content integrity, I want to better understand the results created by our BN feature in a qualitative form. Now that the feature will be out, I want to have a tighter eye on what it delivers. With the Breaking news POC becoming a public demo for feedback, it is not possible for me to look at results at a glance. While cloudwatch is perfectly useful, it is not consumable in that format. I'm seeking to answer a number of questions, but most of all — to track the qualitative KPI - how many accurate results are we delivering? How many false positives are we delivering?
**//The acceptance criteria suggests a spreadsheet - a csv or a google sheet for example - I am open to other consumable formats.//**
**Acceptance Criteria**
- [] Weekly delivery of spreadsheet with these logs to be observed (automated)
- [] New sheet for a new week (for digestibility)
| date_yes | article link | lang | time_between_create_and _yes | # revision of _yes | parameter 1_at_yes | parameter n | data |
| data |
| data |
| data |
| data |
**ToDo**
- [ ] Please let's make sure 100% we both understand what these logs mean! we can re-write them to be clearer
===== Things to consider: =====
* No alarms necessary
* No visualization(grafana) necessary
* [[ https://docs.google.com/document/d/1e0ks55BvXN-RXSfT59sBoySX2WUGam8PviwgiJp-aCE/edit | This doc by Ricardo explains ]] how to use cloudwatch for these logs