Organisations across Africa are hiring data scientists and investing in machine learning platforms, yet many models never reach production or are quietly abandoned because the underlying data was poorly understood. Strong analytics practice, meaning disciplined extraction, quality checks, exploratory work, sound statistics and honest evaluation, is what turns machine learning from an experiment into a dependable decision tool. It is also the skill set that analysts and BI professionals need to move confidently into data science.

The Data and Analytics for Data Science & Machine Learning in Practice Training Course runs over ten days and follows the analytical path of a real project. Days one to five cover framing the business question, writing SQL to assemble modelling tables, cleaning and validating data, exploratory and statistical analysis, and building and evaluating baseline models with scikit-learn. Days six to ten go deeper: advanced feature engineering, imbalanced data and time series, model diagnostics and interpretability, experiment design and causal reasoning, reproducible pipelines with monitoring after deployment, and a capstone project that takes a raw data set through to a presented recommendation.

Every concept is practised in Jupyter notebooks on realistic data from finance, telecoms, health and public services. Participants keep their notebooks, SQL scripts and templates for use in their own work.

Participants who complete the programme will be able to:

• Frame business problems as analytical and machine learning questions with clear success measures.
• Write SQL queries that extract, join and aggregate data into analysis-ready modelling tables.
• Profile, clean and validate data in pandas, documenting quality issues and treatments.
• Apply exploratory analysis and inferential statistics to understand drivers and relationships.
• Engineer features for tabular, categorical, text and time series data while avoiding leakage.
• Evaluate models with appropriate metrics, cross-validation and cost-sensitive threshold selection.
• Design and analyse A/B tests and other experiments to measure real-world impact.
• Communicate model evidence, limitations and recommendations to non-technical decision-makers.

The course is designed for analytically minded professionals who want to strengthen the data side of machine learning, including:

• Data analysts and business intelligence developers moving into data science
• Junior and mid-level data scientists
• Statisticians and economists in research institutions and national statistics offices
• Monitoring, evaluation and research officers working with large data sets
• Risk, credit and fraud analysts in banks, insurers and microfinance institutions
• Actuaries and pricing analysts
• Telecoms, retail and marketing analysts working on churn and customer models
• Health information and epidemiology data officers
• Data engineers who want to understand modelling requirements

Participants complete the programme with:

• A library of reusable SQL and pandas notebooks for extraction, cleaning and profiling.
• A data quality checklist and validation script to run before any modelling project.
• An exploratory analysis report template with standard charts and statistical tests.
• A feature engineering playbook with leakage checks for common data types.
• A model evaluation framework covering metrics, calibration and threshold choice.
• An A/B test design and analysis template with sample size calculations.
• A completed capstone project taken from raw data to a presented recommendation.

The course is taught as a sequence of guided labs, with each concept applied in code the same day. Participants work through:

• Jupyter notebooks prepared for every session, with starter code and worked solutions
• Realistic, anonymised data sets from finance, telecoms, health and public services
• Pair programming and code review to build good analytical habits
• Short statistical clinics that connect theory to the lab results
• Reproducibility practice with version-controlled notebooks and experiment tracking
• A capstone project presented to peers and facilitators

A CPD-accredited Certificate of Completion issued by Vision Reach Global Consultancy is awarded to participants who attend the ten days and submit the lab notebooks and capstone project.

Participants also keep their code, templates and capstone presentation, which form a practical portfolio they can share with employers or use to support continuing professional development records.