Intermediate Pandas Course for Data Cleaning in East Africa & Africa

Course outline and learning expectations

This is a tutor-led course. The outline shows what your tutor will cover; teaching materials and examinations are provided directly to enrolled students.

What you will learn

• Audit unfamiliar datasets and design an intentional data-type and cleaning strategy.
• Normalise messy text, dates, identifiers, categories, and missing-value codes with reusable rules.
• Combine multiple operational tables with cardinality checks, reconciliation, and referential-integrity tests.
• Reshape untidy reports with melt, pivot, stack, unstack, and MultiIndex tools.
• Use advanced groupby, transform, rank, rolling, expanding, and window patterns without unnecessary apply calls.
• Analyse regional time series with resampling, lags, rolling indicators, periods, and timezone-aware timestamps.
• Read from SQL, automate Excel reporting, and build reusable functions for repeated analysis.
• Create data-quality summaries that show what failed, what changed, and what remains uncertain.
• Deliver a portfolio workflow that converts East African operational data into decision-ready outputs.

Modules

1. 1

   Data Auditing and an Intentional Data-Type Strategy

   Profile schema, nulls, duplicates, cardinality, distributions, ranges, and suspicious category values.Use nullable integer, Boolean, string, category, datetime, and timezone-aware types intentionally.Understand Pandas 3.x string behaviour and safe checks that work across modern Pandas versions.Build a data dictionary that records definitions, keys, units, formats, and allowed values.Separate raw, staging, clean, and output data to protect source evidence.
2. 2

   Advanced Cleaning, Text Normalisation and Entity Rules

   Extract structured fields, standardise multiple missing-value codes, and preserve leading zeros.Create rule-based duplicate and entity-resolution checks without hiding uncertain matches.Apply group-aware imputation only when the business meaning supports it.Build transparent outlier flags rather than automatically deleting unusual observations.Record cleaning decisions in functions that can be tested and rerun.
3. 3

   Reliable Joins, Reconciliation and Relational Data

   Design primary, foreign, compound, and surrogate keys for analysis.Use one-to-one, many-to-one, one-to-many, and many-to-many merges deliberately.Apply validate, indicator, suffixes, anti-match checks, and pre-merge key tests.Use concat, update, combine_first, and controlled lookup enrichment.Reconcile row counts, totals, duplicates, and unmatched keys after every important join.Translate familiar SQL JOIN thinking into Pandas workflows.
4. 4

   Reshaping, Tidy Data and MultiIndex

   Recognise wide, long, nested, and presentation-formatted datasets.Use melt, wide_to_long, pivot, pivot_table, stack, and unstack.Work with MultiIndex rows and columns while keeping outputs understandable.Flatten complex column names and preserve business meaning after reshaping.Convert survey grids and monthly cross-tabs into analysis-ready records.
5. 5

   Advanced GroupBy, Transform and Window Analytics

   Use named aggregations, transform, filter, cumulative methods, ranks, shares, and percent change.Distinguish aggregation, transformation, and filtration clearly.Replace avoidable apply calls with built-in vectorised or GroupBy operations.Calculate rolling, expanding, and exponentially weighted indicators.Create within-branch, within-country, cohort, and peer-group comparisons.Handle small groups, missing denominators, and unstable percentage changes responsibly.
6. 6

   East Africa Time-Series Analysis

   Parse mixed timestamps and work with the Africa/Kampala timezone and other regional zones.Use DatetimeIndex, PeriodIndex, date offsets, resample, asfreq, shift, and lag features.Build rolling averages, growth measures, seasonality views, and gap checks.Align irregular events with merge_asof and ordered time data where appropriate.Distinguish missing observations from true zero activity.Create charts that show trends without overstating causality.
7. 7

   SQL, Excel Automation and Reusable Reporting

   Read SQL queries into Pandas and write controlled tables back to SQLite.Parameterise filters and separate database extraction from analysis logic.Read multi-sheet workbooks and produce formatted summary, detail, exception, and metadata sheets.Package repeated steps into small functions with clear inputs and outputs.Create configuration variables for file paths, reporting periods, and country selections.Add logging-style messages, row-count checks, and safe file naming to reporting workflows.
8. 8

   Capstone: East African Sales and Logistics Analytics

   Define operational questions for orders, fulfilment, delivery performance, stock, customers, or branches.Integrate multiple country and reference tables with explicit join controls.Clean and reshape the data, engineer KPIs, and analyse trends and peer performance.Create an executive summary, detail tables, exception outputs, and a documented notebook.Explain data limitations, cross-country comparability, and recommended operational actions.