Clean-room Python case

Python workflow for quick KPI refresh and stakeholder-specific file output.

A public-safe rebuild of the fast-check workflow: refresh latest actuals into a controlled master table, validate KPI status, flag watch items, and split clean stakeholder files by service group and office. All labels and values are synthetic.

Python files

Refresh steps

CSV

Public-safe sample data

Raw company files

Read workflow README Inspect run script

Business problem

Stakeholder KPI files are slow when actuals, totals, and owner slices are updated manually.

The manual workflow creates risk: stale actuals, inconsistent department filters, and late files for department heads or satellite-office owners. This sample shows the Python pattern I use to refresh once, validate centrally, then split clean outputs.

What the code demonstrates

Refresh: updates only the relevant actual fields for the latest month.
Control: checks duplicate keys, missing actuals, and missing owner rules.
Review: calculates KPI achievement and watch-item flags.
Distribution: writes separate files by stakeholder rules.

Workflow

From master KPI table to owner-ready outputs.

Read process README

01Load

Read master template and latest actuals.

02Refresh

Update actual values by month and KPI key.

03Validate

Compare actual to target and flag watch items.

04Split

Create stakeholder files for review owners.

05Log

Write run notes for handoff and QA.

run order

python src/run_pipeline.py

Output:
  refreshed_master.csv
  stakeholder_sales.csv
  stakeholder_office_hub.csv

One orchestrator script calls refresh and split steps so the full handoff can be repeated.

Python files

Small scripts that show the actual refresh logic.

Synthetic data only

refresh_actuals.pyUpdate master KPI actuals split_stakeholder_files.pySplit files by stakeholder rules run_pipeline.pyOrchestrate refresh and split

Generated outputs

What the workflow produces for quick owner review.

refreshed_master.csv

One controlled table with updated actuals, achievement rate, and watch flags.

stakeholder files

Separate clean outputs for service, department, or office owners based on rules.

watch items

Rows below threshold are flagged so reviewers see what needs attention first.

QA checks

Stops duplicate KPI keys and missing stakeholder rules before output files are sent.

Validation

Built for fast review without manual copy-paste.

The sample validates duplicate KPI keys, missing actuals, missing stakeholder rules, and watch items where achievement falls below the review threshold.

Sample inputs