Data Culture Lab

Learn a common skeleton for solving problems with data. Ask where a problem came from, decide whether it is worth solving, fix how the result will be judged before building anything, turn results into decisions and look back. Then learn how that experience builds up into an organization's habits and standards. Every other lab on this site follows this skeleton.

Status: Every unit has a thin first draft (2026-10-30). Case materials and sources will be filled in after the skeleton has been run once through the first application lab
Size: 11 parts, 52 units, each a 60 to 90 minute session
Audience: From decision makers and managers to analysts, engineers and data scientists. It reads without a technical background and points to other labs where techniques are needed
Cases: The fictional holding company danny22.com and its six subsidiaries (retail, manufacturing, finance, energy, mobility and media). The subsidiaries, people and figures are made up and unrelated to real companies. Only the group IT cases are real: they come from the machine that runs this site
Access: culture-lab.danny22.com, open to signed-in readers. Get in touch to request access
Language: The course material is written in Korean

What you learn

Find the problem first: separate the solution from the problem in a request like "build us a model", check that the number that triggered it is right, and count "don't solve it" or "change how the work is done" as successful outcomes.
Decide how to judge first: pick the type of question, then fix the judgment and the decision rule before building. Covers what to do in domains where experiments are hard.
Turn results into decisions: tell "is the model right" apart from "did the decision improve", pass verification and approval by someone other than the builder, and connect the result to a person's or a system's decision.
Look back: blameless postmortems that look at structure rather than people, and judging the quality of a decision rather than its outcome.
Make it stick: people and roles, platforms, governance and risk, maturity assessment and education, so that data culture becomes an organizational habit.

How a unit works

Theory: the concept and the reasons, and where common failures happen in the lifecycle.
Case: the concept applied to a scenario from the subsidiary whose core business is that kind of data.
Workshop: fill in cells of the canvas and check the result against stated criteria.

Course

Part	What you learn	Units
0 What data culture is	Definition and the two loops, why it is hard, the fictional holding company, the skeleton at a glance	5
1 Finding the problem	Background, the value flow, the gate on whether it is worth solving, metric trees, root causes, solutions, baselines	7
2 Data and judgment design	Who made the data and why, measurement design, where ground truth comes from, quality and assets, question types, judgment design, the ladder for judging causality	7
3 Methods and evaluation	The solution and method ladders, is the model right, did the decision improve, reporting uncertainty as ranges	4
4 Decisions and operations	Verification and approval, communicating and spreading, decision records, levels of autonomy, running it as a system, the crisis path	6
5 Looking back	Monitoring and decision tracking, blameless postmortems, concluding and retiring, sharing and standardizing	4
6 People and organization	The people of data culture, roles and collaboration, organizational structure, the data culture team	4
7 Platform	What a platform should do, build or buy, a tour of the group IT platform	3
8 Governance, risk and ethics	Governance starts with records, access and privacy, data sharing between subsidiaries, fairness and the scope of automation, risk gates	5
9 Settling in	Maturity assessment, education and community, from small wins to standards, treating culture building as a problem	4
10 Workshops	A subsidiary case end to end, a problem from your own organization, presenting and looking back	3