Skip to content

← Technology Roadmap

T6–T11 — Python, pandas, Git, Data Formats & Later Tools

These competencies enter only when the DMBOK chapter/lab benefits from them. The program does not front-load every tool.

T6 — Python Foundations for Data Management

Why: Python supports repeatable file handling, validation, small-scale automation and inspection when SQL alone is not the best fit.

Learn

  • Variables and basic data types.
  • Lists/dictionaries at practical depth.
  • Conditions and loops.
  • Functions.
  • Reading/writing files.
  • Modules/imports and virtual-environment awareness.
  • Exceptions at a basic practical level.
  • Clear script input/output and logging/print evidence.

Governance/Data Management uses

  • Validate incoming files.
  • Compare metadata or reference lists.
  • Generate/reproduce synthetic lab data.
  • Automate bounded evidence collection.
  • Call APIs later when a governance/catalog scenario requires it.

Readiness proof

Write a small script from scratch that reads a Meridian input, applies a transparent business/data rule, reports exceptions, and writes a separate output without overwriting the raw source.

Stop boundary

No web frameworks, application architecture, algorithm-heavy CS or software-engineering specialization.

T7 — pandas for Data Management

Why: pandas makes tabular profiling, cleansing, reconciliation and source comparison visible without requiring a production pipeline.

Learn

  • DataFrame/Series concepts.
  • Load/save CSV and other relevant tabular formats.
  • Inspect shape, columns, types and sample values.
  • Filter/sort/select.
  • Null detection and basic profiling.
  • Duplicate detection.
  • Grouping/aggregation.
  • Merge/join at practical depth.
  • Controlled type/string/date cleanup.
  • Compare/reconcile datasets.

Governance/Data Management uses

  • Profile Customer source conflicts.
  • Detect invalid codes or missing required values.
  • Build before/after quality evidence.
  • Reconcile file and database extracts.
  • Prepare bounded metadata or catalog imports.

Readiness proof

Profile two synthetic Meridian customer sources, identify null/duplicate/format differences, join or reconcile them appropriately, and explain the business meaning of the findings.

Stop boundary

No ML/scientific-computing specialization.

T8 — Python ↔ PostgreSQL Bridge

Why: Some Meridian labs need repeatable movement between file/Python analysis and the relational lab database.

Learn

  • Connect using a supported Python database library/SQLAlchemy pattern at a basic level.
  • Read a query result into Python/pandas.
  • Write a bounded DataFrame/table output intentionally.
  • Parameterized query awareness.
  • Transaction/error awareness.
  • Keep credentials outside committed source.

Readiness proof

Read a Meridian table into pandas, perform one transparent validation/transformation, write a clearly named lab output or evidence table, and verify the result with SQL.

Stop boundary

No production ingestion framework, orchestration, connection-pool engineering or pipeline platform design.

T9 — Git Fundamentals

Why: Data Governance and Data Management work benefits from reproducible, reviewable changes to SQL, scripts, definitions, decision records and lab evidence.

Learn

  • Repository, working tree, staging, commit.
  • status, add, commit, basic history/diff.
  • Meaningful commit messages.
  • Basic branches only when useful.
  • .gitignore and secrets awareness.
  • Difference between source/evidence and generated/disposable artifacts.

Readiness proof

Create or update a lab file, inspect the diff, commit it with a meaningful message, then explain what changed and why.

Stop boundary

No advanced Git administration, CI/CD or complex branching strategy required.

T10 — Data Formats: CSV, JSON & Parquet

Why: A governance/data-management practitioner needs to recognize how format affects structure, metadata, quality checks, lineage and exchange.

CSV

Know headers, delimiters, quoting, encoding/type ambiguity, missing values and why schema is often external/implicit.

JSON

Know objects, arrays, nested structures, keys/values, schema drift and why nested data can complicate mapping/profiling.

Parquet

Know columnar file concept, embedded schema/metadata at practical level, analytics-oriented strengths and why it differs from row-oriented text formats. Chapter 14 uses it for format/scale awareness.

Readiness proof

Inspect the same or related Meridian business data in at least two formats and explain the governance/quality/metadata implications of each.

Stop boundary

No distributed-storage engineering required.

T11 — Chapter-Triggered Later Tools

Metabase — Chapter 11

Use to make warehouse/BI outputs observable: connect to the lab database, create a small number of governed questions/measures/dashboard views, and trace displayed results back to definitions/data. Stop before BI-platform administration.

OpenMetadata — Chapter 12/13 optional

Use only after the learner manually understands glossary, dictionary, catalog, ownership, classification and lineage. It is a sandbox for seeing metadata-governance concepts in a platform, not the source of those concepts. Stop before platform-administration specialization.

DuckDB — Chapter 14

Use for local queries against larger CSV/JSON/Parquet/event data so format/scale differences can be explored without a distributed-engineering stack. Stop before large-scale platform engineering.

REST APIs / Postman — only when a Chapter 8 or later lab needs them

Learn request/response, endpoint, method, parameters, headers, JSON payload, status code and authentication concepts at API-literacy depth. Use Postman or an equivalent client to inspect/test a bounded synthetic/official endpoint. Stop before API application development.

Timing rule

These tools are chapter-triggered. Do not learn Metabase, OpenMetadata, DuckDB or Postman early simply because they are in the roadmap.

Source: Technology Competency Roadmap; Detailed Technology Competency Reference; 00C technology-gate logic.