nbadbArena Data Lab
GuidesSet MenuRole Based Onboarding Hub15 waypoints

Playbook

Role-Based Onboarding Hub

Coaching-staff hub for choosing the right nbadb route by role, task, and first command.

Role-Based Onboarding Hub

Welcome to the coaching staff room for nbadb. Use this page to pick your role, run the opening drill, and get onto the right page before you burn time wandering the whole playbook.

Quick route board

RoleFirst readFirst commandNext stop
ContributorInstallationuv sync --extra devCLI Reference
AnalystGetting Startednbadb download or open nba.duckdbAnalytics Quickstart
Data engineerDaily Updatesnbadb dailyTroubleshooting Playbook
StakeholderArchitecturenbadb schemaSchema Reference

If you are not sure where to begin, start with the row that matches the job you need done today, not the title on your org chart.

Pick your lane

Each tab gives you the same scan path:

  1. Run first to get moving
  2. Open next so you land on the right follow-up page
  3. Watch for so the common misses do not cost a second possession

Primary route: InstallationCLI ReferenceArchitecture

Run first

  1. Set up the repo-local environment:
    uv sync --extra dev
  2. Keep the current command surface nearby in CLI Reference.
  3. Read Architecture before changing pipeline flow, table families, or generated docs.

Open next

RouteWhy
Schema ReferenceUnderstand output naming and table families
Data LineageTrace a transform back upstream
uv run nbadb docs-autogen --docs-root docs/content/docsRegenerate generator-owned docs after schema changes

Watch for

  • Generator-owned docs should be regenerated, not hand-edited.
  • Underscore-prefixed pipeline tables are internal operational state, not part of the public analytical contract.
  • If CI or artifacts fail, jump to the Troubleshooting Playbook.

Primary route: Getting StartedAnalytics QuickstartDuckDB Query Examples

Run first

  1. Seed a local dataset:
    nbadb download
    or use an existing nba.duckdb / Parquet export.
  2. Start with analytics_* surfaces for the fastest first query.
  3. Move into the table-level docs only after you know the question you want to answer.

Open next

RouteWhy
DuckDB Query ExamplesSQL-first scouting packet
Player ComparisonHead-to-head analysis patterns
Shot Chart AnalysisCourt-map and zone work
Parquet UsagePolars, Pandas, and Arrow workflows

Watch for

  • Keep season_year and season_type explicit so repeated analysis stays comparable.
  • DuckDB is the fastest first possession for joins; Parquet is the cleaner handoff when the rest of the workflow is DataFrame-first.

Primary route: Daily UpdatesCLI ReferenceKaggle Setup

Run first

  1. Review what each run mode actually does: daily, monthly, full, status, run-quality.
  2. Confirm your data directory and output expectations.
  3. Keep the latest health artifacts in view:
    • artifacts/health/daily/dependency-inventory.json
    • artifacts/health/daily/data-quality-report.json
    • artifacts/endpoint-coverage/*

Open next

RouteWhy
Pipeline FlowShare a stage map quickly
ArchitectureReview validation tiers, internal state, and run-mode boundaries
Kaggle SetupHandle dataset handoff and publish flow

Watch for

  • daily, monthly, and full end by rebuilding downstream tables in replace mode.
  • A single Ctrl+C is the graceful stop; a second interrupt forces exit.
  • When something breaks, start with the exact artifact or command named in the failure message.

Primary route: ArchitectureSchema ReferenceER Diagram

Run first

  1. Read Architecture for the source-to-export story.
  2. Scan Schema Reference to understand what the public dataset actually contains.
  3. Review Kaggle Setup when the conversation shifts to distribution and delivery.

Open next

RouteWhy
ER DiagramFast visual schema read
Endpoint MapSource-coverage conversations
Strategic Shift RolloutPlanned migration sequencing

Watch for

  • The public analytical contract is the star/analytics surface, not the underscore-prefixed internal pipeline tables.
  • Use this hub as the handoff page for new teammates entering the rotation.

Keep moving

Stay in the same possession

Keep the mental model warm with adjacent pages, section hubs, and search-friendly routes into the same topic cluster.

Section hub

On this page