Skip to main content

Topic roadmap

Data

11 lessons, about 37 hours in all, from the basics to advanced. Start with the first, or jump to what you need.

Lessons, easiest first

  1. Data · Beginner 2 · about 2 hReading and shaping JSONRead nested JSON from APIs and webhooks, pick out the fields you need, reshape lists and objects, and handle missing values, in an automation tool without heavy code.
  2. Data · Beginner 3 · about 2 hData strategy for AI productsPlan the data an AI product needs: what, from where, with what consent and quality, how it is labelled and evaluated, and how user feedback becomes better data over time.
  3. Data · Beginner 3 · about 4 hFiles, tables and PDFsRead the messy files clients actually send: CSV and Excel with odd encodings, nested JSON, and PDFs with text, tables and scans.
  4. Data · Beginner 3 · about 4 hNumPy and plottingWork with numerical data in NumPy (arrays, shapes, vectorised maths, broadcasting) and explore it visually with Matplotlib histograms, scatter plots and line charts.
  5. Data · Beginner 3 · about 5 hSQL for engineersAnswer real business questions in SQL with joins, GROUP BY, window functions and indexes, and read a slow query's plan.
  6. Data · Beginner 3 · about 4 hTables with pandasLoad a CSV into pandas, then select, filter, group and sort the data.
  7. Data · Intermediate 2 · about 3 hData quality and personal dataCheck data before it reaches the model, and find and mask personal data such as names, phone numbers, Aadhaar and PAN numbers in documents, prompts and logs.
  8. Data · Intermediate 2 · about 2 hData versioningVersion datasets and data pipelines like code so every model can be traced to the exact data it was trained on, and any training run can be reproduced.
  9. Data · Intermediate 2 · about 4 hParsing documents and OCRTurn PDFs, scans, slides and Word files into clean, ordered text and intact tables for retrieval, choosing between text extraction, OCR and layout-aware parsing.
  10. Data · Pro 1 · about 3 hBuilding a training dataset: quality, coverage and quantityPlan, write, check and size a dataset for fine-tuning: the right format, a labelling guide people agree on, enough variety, and a test that shows when more data will help.
  11. Data · Pro 1 · about 4 hData pipelines for AIDesign the pipeline that keeps an AI system's data fresh: ETL or ELT, scheduled or change-driven updates, incremental indexing, and re-processing when parsers or embedding models change.

On these role roadmaps

Interview questions on data

Books on data

All recommended books

Today's AI turns are used

Lessons, roadmaps, videos, practice cards and the code playground stay free with your account, with no limit.