Day 1 – AI in Data Analysis
Translating business requirements into analytical questions, hypotheses and
KPIs
Understanding the data model and selecting relevant variables with AI
support
Designing an analytical plan and workflow in Python or
SQL
Exploratory data analysis with Pandas and AI: descriptive statistics,
distributions, correlations, segmentation, time series and outliers
Verifying
AI-suggested hypotheses with calculations in Python
Hands-on exercise: from a
business question through the analysis of an unfamiliar dataset to a verified
conclusion
Day 2 – AI in Data Preparation and Data Quality
Checking for missing values, categories, duplicates, data types, extreme
values and dates
Designing and validating data quality rules and business
rules
Designing a transformation workflow with AI and implementing it in
Python
Reviewing generated code: preserving the meaning of data, handling NaN
values and outliers, and avoiding the risk of data leakage
Combining CSV,
Excel, SQL and API sources; checking schemas, keys and joins
Hands-on Dirty
Data Challenge: building a reproducible data-cleaning workflow
Day 3 – Exploratory Analysis, Visualisation and Storytelling
Finding trends, seasonality, segments, correlations, anomalies and changes in
behaviour
Distinguishing between facts, findings, hypotheses and
interpretations; validating analytical conclusions
Choosing a visualisation
to test a specific hypothesis
Generating and reviewing visualisation code in
Plotly or Matplotlib
Translating findings into business impacts and
recommendations for management
Hands-on exercise: five findings, three
visualisations, an executive summary and recommendations, with facts, AI
interpretations and hypotheses clearly labelled
Day 4 – AI in Machine Learning
Designing an ML experiment: target variable, input features, preprocessing,
algorithms, validation strategy and metrics
Identifying data leakage and
reviewing the experiment design
Building and reviewing a Python pipeline for
data preparation, feature engineering, model training and
evaluation
Comparing Logistic Regression, Random Forest and XGBoost
models
Interpreting accuracy, precision, recall, F1 score, ROC AUC, RMSE, MAE
and the confusion matrix
Explaining models with SHAP and using LLMs to
formulate business explanations
Hands-on exercise: a complete AI-assisted ML
experiment, from the business problem to a recommendation
Day 5 – AI Agent in the Work of a Data Analyst
Delegating parts of the analytical work to an agent with access to Pandas,
SQL, ML and visualisations
A multi-step workflow: from data exploration
through Python or SQL and hypothesis testing to a final report
Analyst
checkpoints and evidence-based decision-making
Working with data securely:
cloud LLMs, personal data, trade secrets, anonymisation and
pseudonymisation
Reviewing AI-generated code, auditability, reproducibility,
hallucinated conclusions, data leakage and prompt injection through external
data
Final project: analysing production data, downtime, quality and costs;
identifying problems, building a predictive model and proposing corrective
measures
Combining Python, Pandas, SQL, visualisations, ML, LLMs and an
analytical AI agent in a single project
