New: Roadmaps ordered paths through our cheat sheets and flashcards, so you always know what to study next.
Explore themSee what's new on GitHubFrom a star schema and DAX to a governed, secured BI platform the whole company runs on.
A 13-step learning path. Follow it in order, or jump to what you need.
This path is for analysts who already have a dashboard or two under their belt and want to own BI for a whole organization: modeling data correctly, securing it, governing it, and making sure it performs at scale rather than building one more personal report. Plan on about 7 to 12 weeks at a few hours a week, moving from dimensional modeling and DAX through semantic layers, security, governance, and performance tuning to dashboard design and decision intelligence. It does not start you from zero on Excel or Power BI, that ground belongs to the Data Analyst and Excel to Power BI paths, and it stays light on the data engineering pipelines that feed a warehouse, which belong to the Data Engineer path. By the end you can design a governed semantic model with proper security, diagnose and fix a slow BI report, and turn a dashboard into a decision the business actually commits to.
Expected: basic Excel and a first Power BI report, formulas, a pivot table, and a published dashboard. Helpful but not required: SQL.
Open this first: you likely already know one spreadsheet or one Power BI report, and this cheat sheet hands you the enterprise vocabulary, from data warehouse to OLAP cube to semantic layer, that every step from here on assumes you already have.
Table 2's modeling approaches from step 1 turn concrete here: fact tables, dimension tables, and the grain decision that determines whether every query built on top of this model tells the truth.
A star schema from step 2 is just tables until you can calculate against it, and CALCULATE's filter context is the concept the field says trips up more BI learners than anything else: expect to come back to it more than once before it clicks.
You can design a star schema with the grain declared before anything else, and you can write a DAX measure that recalculates itself instead of freezing into a static number. Next up: making sure only the right people can see what you just built.
Finish this section to unlock.
+100 XP
DAX in step 3 defines a measure inside one report; a semantic layer is how an enterprise keeps that same measure honest everywhere it gets reused, from dashboards to the AI agents now querying it directly.
Every measure in step 4's semantic layer is worthless if the wrong person can see it, and dynamic row-level security is where filter propagation gets genuinely confusing: expect to test it more than once before you trust it in production.
Row-level security in step 5 locks down one control; a real governance program is the roles, lineage, and policy backbone that decides who owns that control and how it gets enforced across the whole company.
Step 6's governance framework exists for exactly this moment: the safe boundaries that let you hand business users their own reports without the shadow-spreadsheet chaos self-service usually invites.
You can lock a semantic model down with row-level security, back it with a real governance program, and still hand safe self-service access to business users instead of hoarding every report; a few minutes of due flashcards keeps step 2's star schema fresh while you build on it. Next up: making all of it fast and demonstrably reliable.
Finish this section to unlock.
+100 XP
Choose this route if you're the one signing off on data-quality SLAs, not just trusting someone else's clean export: the profiling and monitoring discipline that keeps step 6's governance program honest over time.
Come back to step 2's schema and step 3's DAX with a harder question: not whether they work, but whether they still work once the model hits fifty million rows.
Pick this option if your organization runs Analysis Services or another OLAP cube engine instead of a Power BI tabular model: the query language that plays the same role there as DAX has played since step 3.
You can trace a slow report back to its indexing and DAX filter context, know when a formal data-quality program earns its keep, and reach for MDX when the platform is an OLAP cube instead of a tabular model. Next up: putting it all in front of people and turning it into decisions.
Finish this section to unlock.
+100 XP
All the modeling, security, and speed work from steps 2 through 9 still needs a screen a director actually reads; this is the layout and color discipline that makes a KPI register in five seconds instead of getting ignored.
Consider this if last night's refresh behind step 11's dashboard isn't fast enough for the decisions your business needs to make right now: streaming pipelines and low-latency delivery in place of batch loads.
Pull together the model from step 2, the governance from step 6, and the dashboard from step 11 into their actual point: turning insight into a decision someone in the business commits to, the enterprise BI loop finally closing.
You can model, secure, govern, and tune a BI platform, then design a dashboard people actually read and turn what it shows into a decision the business commits to, the complete loop the BI Analyst badge certifies.
Finish this section to unlock.
+100 XP
Finish every required step, at least 70% of them genuinely done (not skipped), to earn this badge and 500 XP.