Lesson 1Deployment considerations: retraining cadence, data pipelines, monitoring drift and backtesting forecastsPlan deployment of revenue forecasting systems, including retraining schedules, automated data pipelines, monitoring for data and concept drift, and continuous backtesting to ensure steady performance in changing market conditions.
Designing robust data pipelinesScheduling retraining and updatesMonitoring data and concept driftOngoing backtesting and benchmarkingAlerting and rollback strategiesLesson 2Communicating forecasts to stakeholders: visualization of point forecasts and intervals, scenario analysis, transparency of assumptionsShare revenue forecasts effectively with stakeholders using clear visuals of point and interval predictions, scenario analysis, and open documentation of assumptions, limitations, and model risk considerations.
Visualizing point and interval forecastsScenario and what-if analysis designExplaining drivers and key featuresDocumenting assumptions and limitsTailoring messages to stakeholdersLesson 3Machine learning time series models: random forests/gradient boosting with lagged features, XGBoost/LightGBM, and sequence models (LSTM/GRU)Train machine learning models for time series revenue forecasting, including tree ensembles with lagged features and sequence models like LSTM and GRU, while handling nonstationarity, seasonality, and product-level differences.
Random forests with lagged featuresGradient boosting, XGBoost, LightGBMGlobal versus local forecasting modelsSequence models with LSTM and GRUHandling nonstationarity and scalingLesson 4Formulating forecasting objectives and evaluation horizons (e.g., next 3, 6, 12 months)Define forecasting goals for core bank products by choosing prediction targets, horizons, and detail level, and align them with business decisions such as budgeting, pricing, liquidity planning, and regulatory or risk reporting needs.
Choosing revenue targets and unitsSelecting forecast horizons and frequencyAligning forecasts with business decisionsGranularity by product, segment, and regionHandling new products and short historiesLesson 5Identifying and sourcing time series data (public financial series, payment volumes, synthetic generation techniques)Learn how to identify, assess, and source time series data for bank revenue forecasting, including internal product metrics, public financial series, and synthetic data that safely boosts scarce or noisy historical records.
Cataloging internal product revenue seriesUsing public macro and market data sourcesCollecting payment and transaction volume dataAssessing data quality, gaps, and revisionsSynthetic data generation for stress scenariosLesson 6Training and hyperparameter tuning: grid/random search, Bayesian optimization, time-aware scoringOptimise model performance using structured hyperparameter tuning strategies, including grid and random search, Bayesian optimisation, and time-aware scoring that respects temporal order and focuses on business-critical horizons.
Defining search spaces and priorsGrid and random search trade-offsBayesian optimization workflowsTime-aware validation and scoringEarly stopping and resource limitsLesson 7Model ensembling and reconciliation: simple model averaging, weighted ensembles, stacking for time seriesCombine multiple forecasting models for bank revenue using simple averages, weighted ensembles, and stacking, and apply hierarchical reconciliation to ensure coherent forecasts across products, branches, and organisational levels.
Simple and weighted model averagingStacking and meta-learners for seriesDiversity and correlation among modelsHierarchical and grouped reconciliationEvaluating ensemble stability over timeLesson 8Baseline time series methods: ARIMA, ETS, naive and seasonal naïve models, decomposition (trend/seasonality)Explore baseline time series models for bank revenue, including naive, seasonal naive, ARIMA, ETS, and decomposition, to establish reference performance and interpret trend and seasonality before using complex machine learning models.
Naive and seasonal naive benchmarksClassical decomposition of trend and seasonalityARIMA modeling for bank revenue seriesExponential smoothing and ETS variantsComparing baselines across productsLesson 9Feature engineering for revenue: lags, rolling means/std, differencing, calendar effects, holiday indicators, cohort effects, marketing/campaign flagsEngineer predictive features for bank revenue, including lags, rolling statistics, differencing, calendar and holiday effects, cohort and lifecycle indicators, and marketing or campaign flags that capture demand shifts and structural breaks.
Lag and lead features for revenueRolling means, volatility, and ratiosCalendar, holiday, and payday effectsCohort and lifecycle based featuresMarketing and campaign impact flagsLesson 10Data splitting and cross-validation for time series: train/validation/test splits, expanding window CV, blocked CVDesign time-aware data splits and cross-validation schemes for revenue forecasting, including rolling and blocked approaches, to avoid leakage, mimic production use, and obtain reliable estimates of model performance over time.
Holdout train, validation, and test splitsRolling and expanding window validationBlocked cross-validation for seasonalityPreventing temporal leakage in featuresBacktesting over multiple forecast originsLesson 11Evaluation metrics and error analysis: MAE, RMSE, MAPE, symmetric MAPE, prediction intervals and coverageEvaluate revenue forecasts using metrics such as MAE, RMSE, MAPE, symmetric MAPE, and interval coverage, and perform detailed error analysis by segment, horizon, and regime to uncover biases and model weaknesses.
Scale-dependent error metricsPercentage and relative error metricsPrediction intervals and coverageHorizon and segment level diagnosticsRegime and event-driven error analysisLesson 12Macro and external regressors: using CPI, unemployment, interest rates, mobility, Google Trends; feature selection and lag alignmentIncorporate macroeconomic and external regressors into revenue models, such as CPI, unemployment, interest rates, mobility, and search trends, and learn techniques for lag alignment, scaling, and feature selection to avoid overfitting.
Selecting relevant macro indicatorsAligning lags between macro and revenueTransforming and scaling external dataFeature selection and regularizationStress and scenario overlays with macros
