Free to read ยท Three levels
Data Analysis: Questions, Evidence, and Reproducible Results
Turn synthetic tables into defensible summaries, visual explanations, and carefully qualified decisions.
Start learningIn this course
This complete theory course starts with analytical questions, measurement units, CSV ingestion, and data quality. It develops grouping, SQL joins, visualization, sampling, uncertainty, experiments, time series, reproducibility, and responsible communication across eighteen substantial lessons.
Examples use small invented records, Python, and isolated in-memory SQL. Worked calculations expose assumptions and denominators instead of treating a chart or a statistical function as an answer by itself. Every lesson includes diagnosis guidance and solved exercises; the Arabic edition preserves the same full concepts and examples.
What you will learn
- Define a population, observation unit, metric, and data contract.
- Clean, join, group, and visualize data without silently changing its meaning.
- Separate descriptive patterns from causal and statistical claims.
- Create reproducible analyses and communicate uncertainty and limitations.
Your learning path
Three levels
Basic: Questions and Trustworthy Tables
Define observations and build auditable summaries from typed, validated records.
- Questions, Grain, and Provenance
- Data Types, Measurement, and CSV
- Cleaning and Validation Rules
- Missing Values and Duplicate Records
- Descriptive Statistics and Distributions
- Grouping, Rates, and Denominators
Intermediate: Relationships and Uncertainty
Combine tables, inspect patterns, and relate sampling design to uncertain estimates.
- SQL Query Logic and Honest Denominators
- Joins, Cardinality, and Changing Grain
- Visualization Design, Bins, and Uncertainty
- Exploratory Relationships and Stratification
- Sampling Frames, Selection, and Weights
- Standard Errors, Confidence Intervals, and Assumptions
Advanced: Defensible and Reproducible Decisions
Handle causality, time dependence, resampling, pipelines, and communication with explicit limitations.
- Comparisons, Experiments, and Causal Claims
- Time Series, Windows, and Leakage
- Resampling, Tests, and Multiple Comparisons
- Reproducible Analytical Pipelines
- Scale and Numerical Reliability
- Communication and Responsible Decisions
Who this is for
- Learners who want a careful introduction to practical data analysis.
- Developers and analysts reviewing data quality, metric definitions, uncertainty, and reproducibility.
Before you start
- Basic arithmetic and comfort reading small tables.
- Some Python familiarity helps, but examples explain the analytical role of each operation.
- Complete preceding levels or bring equivalent knowledge before advanced inference topics.