Python Data Science and Predictive Modelling Training Course

Course overview

Many teams collect operational, customer, financial and digital data but still rely on spreadsheet summaries, manual extracts or black-box dashboards when they need to explain what is happening and predict what will happen next. This course equips professionals to use Python to turn raw tabular data into reliable analysis and defensible predictive models. Participants learn how to identify a business question that data can answer, prepare data without obscuring quality issues, and communicate findings in terms that decision-makers can act on.

Across five days, participants work with Python, JupyterLab, pandas and scikit-learn to import, inspect, clean, reshape and visualise datasets. They build reproducible analysis notebooks; calculate descriptive statistics; engineer meaningful features; and develop regression and classification models. The programme covers train/test splits, cross-validation, model metrics, baseline comparisons, overfitting, parameter tuning and interpretation. Participants also learn to select appropriate measures such as RMSE, MAE, accuracy, precision, recall, F1 score and ROC-AUC according to the cost of different business errors.

Teaching combines concise instructor demonstrations with guided coding labs, practical case data and structured review of results. Rather than only running pre-written code, participants make modelling choices, diagnose data problems and explain their choices to peers. By the end of the course, each participant leaves with a documented Jupyter notebook containing a complete predictive modelling workflow: data preparation, exploratory analysis, feature engineering, model evaluation, visualisations and a short set of business recommendations.

The course is suited to analysts, technically confident business professionals and early-career data practitioners who need a practical route from data files to usable predictive insight. It is also valuable for managers commissioning analytics work who need to assess model outputs, assumptions and limitations more effectively.

Course objectives

By the end of this course, participants will be able to:

• Import and profile CSV and Excel data using Python and pandas DataFrames
• Clean missing, duplicated and inconsistent records using reproducible pandas transformations
• Create exploratory visualisations and summary statistics that answer defined business questions
• Engineer numeric, categorical and date-derived features for predictive modelling
• Build baseline and improved regression models with scikit-learn pipelines
• Train classification models and evaluate them using precision, recall, F1 score and ROC-AUC
• Apply train/test splits, cross-validation and parameter tuning to control overfitting
• Produce a documented Jupyter notebook that explains model results and business recommendations

Benefits of attending

Target competencies

Who should attend

• Data Analysts — who need to move from descriptive reporting to reproducible predictive analysis
• Business Intelligence Analysts — who need to investigate drivers and forecast outcomes beyond dashboard metrics
• Operations Analysts — who need to predict demand, delays, failures or service outcomes from operational data
• Marketing Analysts — who need to segment customers and model conversion, churn or campaign response
• Finance Analysts — who need to build evidence-based forecasts and risk classifications from tabular data
• Digital Product Managers — who need to interpret model performance and commission data science work intelligently

Requirements and prerequisites

Participants should be comfortable using a computer, working with files and interpreting tables, percentages and basic charts. Familiarity with spreadsheet formulas and basic descriptive statistics, such as averages and distributions, is helpful. No prior data science, machine learning or scikit-learn experience is required. Some exposure to Python variables, lists, loops or functions is beneficial but not essential; the course introduces the Python syntax needed for analysis. Complete beginners should expect a coding-intensive week and should be prepared to practise short exercises daily. Participants need access to a laptop capable of running Python and JupyterLab.

Training methodology

The instructor introduces each technique through live Python demonstrations in JupyterLab, then participants apply it to structured business datasets in guided labs. Exercises progress from data inspection and cleaning to visual analysis, feature construction and model comparison using pandas and scikit-learn. Short case discussions require participants to choose metrics based on business consequences, not just technical scores. Peer review is used to challenge assumptions and interpret outputs. On the final day, participants complete and present a predictive modelling notebook with an application plan for a relevant workplace dataset.

Course outline

Day 1: Python foundations for analytical data work

• JupyterLab notebooks, kernels and reproducible cell execution
• Python variables, strings, lists, dictionaries and control flow
• NumPy arrays and vectorised numerical operations
• pandas Series and DataFrame structures
• Importing CSV and Excel files with pandas
• Data types, indexing, filtering and sorting
• Initial data profiling with info, describe and value counts

Workshop: Participants import an operational dataset into a Jupyter notebook and produce a data-profile sheet identifying field types, missingness and initial data-quality questions.

Day 2: Data preparation and exploratory analysis

• Missing-value diagnosis and imputation strategies
• Duplicate detection and record-level data-quality checks
• String cleaning, category standardisation and type conversion
• Merging, concatenating and reshaping DataFrames
• Groupby aggregation and pivot-table analysis
• Distribution analysis, outlier identification and skewness
• Visualisation with matplotlib and pandas plotting

Workshop: Participants clean and combine source files, then create an exploratory analysis notebook with charts and a prioritised list of data-quality decisions.

Day 3: Features, targets and regression modelling

• Defining prediction targets and measurable business outcomes
• Feature engineering from numeric, categorical and date fields
• One-hot encoding and scaling within preprocessing pipelines
• Train/test splits and data leakage prevention
• Linear regression and regularised regression models
• Regression baselines, residuals and error analysis
• MAE, RMSE and R-squared metric selection

Workshop: Participants build and compare baseline and regression models to predict a continuous business outcome, documenting the selected metric and model rationale.

Day 4: Classification and model improvement

• Binary classification use cases and class labels
• Logistic regression and decision tree classifiers
• Confusion matrices and threshold-based decisions
• Accuracy, precision, recall, F1 score and ROC-AUC
• Class imbalance and stratified sampling
• Cross-validation and GridSearchCV parameter tuning
• Overfitting, underfitting and learning from validation results

Workshop: Participants create a classification model for a churn or risk case, compare thresholds and recommend an operating point based on the cost of false positives and false negatives.

Day 5: Model interpretation and workplace application

• Feature importance and coefficient interpretation
• Partial dependence concepts and practical model explainability
• Prediction pipelines for repeatable scoring
• Model limitations, bias considerations and monitoring requirements
• Communicating uncertainty and avoiding misleading claims
• Structuring an analytical notebook for stakeholder review
• Planning a workplace predictive modelling use case

Workshop: Participants complete and present an end-to-end predictive modelling notebook, including recommendations, limitations and a practical plan for applying the workflow to a workplace problem.

Tools & standards covered

Python, JupyterLab, pandas, scikit-learn

A typical training day

Live online deliveries follow the same structure in the East Africa Time zone, with shorter screen blocks and longer breaks.

What the fee includes

How you can take this course

Certification

Participants who complete the full five days receive the Skillset Development Certificate of Completion, stating the course title, course code, dates and delivery format — suitable for professional-development records and employer reimbursement.

Frequently asked questions