How to take the Challenge
This is a project-controls mapping problem, not a coding exam. Use the tools and judgment you would actually use.
What you're trying to do
You have an owner WBS, schedule activities, cost information and supporting contractor/fabricator records that describe the same project in different ways. Your job is to build a reviewable crosswalk back to the owner WBS.
1. Start with Practice
The Practice set includes its reference mapping. Try your approach, run the local scorer and see where your logic differs before taking the hidden Challenge.
Excel, Power Query, SQL, Python, AI, or a combination are all fair game.
2. Map each record
Return the applicable WBS Cell ID, or use one of these plain-language outcomes:
If one record genuinely spans two WBS cells, use A|B.
3. Add confidence
Use high, med or low for answered mappings. A confident wrong mapping is penalized most heavily because it can pass through a reporting process without being challenged.
A missing mapping creates a question. A confident wrong mapping creates a reporting risk.
4. Score Practice locally
From the extracted Practice folder, run:
python scorer/score.py my_answers.csv --set practice --exceptions my_exceptions.csv
The Practice set is repeatable. The hidden Challenge allows one scored attempt per GitHub user.
5. Submit, then tell us what we're missing
Your score is only part of the point. The submission form includes a short professional debrief about the evidence you relied on, how mapping risk should be treated, what makes this harder on real projects, and how much effort reconciliation takes today.
Please keep examples general. Never post real client, employer or confidential project information.
Where this is going
WBS↔LINK is an early-stage concept for a secure, AI-enabled project-controls mapping workflow: connect schedule, cost, WBS and contractor records; propose mappings; surface uncertainty for review; reuse approved relationships in later reporting cycles; and export a standard Excel/CSV crosswalk for teams that want to continue working in their existing tools.
The public challenge is how we pressure-test the problem before pretending the product is finished.