Weekly status update:

Cleaned up and formatted test dataset. Identified and created benchmarks for some data elements, and revisited how to implement weights and scoring of individual elements for metrics computation. Completed 80% of analysis and report generation for aggregate statistics (for the entire test dataset). Next week: finish analysis/report for aggregate dataset and test metrics computation. Start on collection-level analysis and report generation.

Weekly status update:

• Created basic python scripts to import test data from CSV, calculate metrics from the data, and output scores.
• Implemented the first metrics calculation function for processing test data, and confirmed that the basic logic is working. Next week I will be able to more quickly implement functions for all the other calculations now that I have a working example.

Weekly status update:

• Finished implementation of python scripts to output calculations for all test metrics.
• Generated metrics output for the test dataset: raw output is computed based on 30 doc attributes across 140 pages, for 7 metrics categories. (raw output data for test set of pages is attached to this task as a csv)
• Added doc collections to raw output to enable analysis of metrics scores for each of the 5 collections in our test set.
• Began analysis of the test metrics output!

Next week: Finish intitial analysis and discuss {F58325103}outcomes with Tech Docs team. Stretch: create a plan for gathering feedback from community and stakeholders.

A version of the metrics test output with an additional field indicating each pages's collection membership is now available at https://gitlab.wikimedia.org/repos/technical-documentation/doc-metrics-testing/-/blob/develop/data/full_output%20with%20collections.csv

I have now published comprehensive reference documentation covering the fields in our test dataset (the metrics input), and the fields in the metrics output: https://www.mediawiki.org/wiki/Wikimedia_Technical_Documentation_Team/Doc_metrics/Field_reference

Next week I will be working with the other technical writers to try out some persona-based testing of the metrics output, and writing a short-and-sweet user guide (see: draft). When that is done, we'll be ready to share all this work more widely and formally request feedback from the larger community and stakeholders.