Automobile demand forecasting: Spatiotemporal and hierarchical modeling, life cycle dynamics, and user-generated online information

TL;DR

The paper tackles monthly automobile demand forecasting for a German premium OEM across a multi-product, multi-market hierarchy, addressing data sparsity, life-cycle effects, and the value of online information. It extends LightGBM-based probabilistic forecasting with data-driven pooling (POOL-SEL-BP) and an integer-coherent reconciliation (REC-MILP/REC-LW-MILP), plus a life-cycle AVM feature, to produce coherent point and probabilistic forecasts across levels. SHAP analysis identifies dynamic short- and mid-term drivers, and linear mixed-effects models show that online configurator data improves disaggregated forecasts, with gains attenuating at higher levels. The results demonstrate context-dependent forecasting gains, validate the need for integer-consistent reconciliation for operational feasibility, and offer actionable guidance for deploying pooled ensembles and online-information signals in industrial hierarchical forecasting.

Abstract

Premium automotive manufacturers face increasingly complex forecasting challenges due to high product variety, sparse variant-level data, and volatile market dynamics. This study addresses monthly automobile demand forecasting across a multi-product, multi-market, and multi-level hierarchy using data from a German premium manufacturer. The methodology combines point and probabilistic forecasts across strategic and operational planning levels, leveraging ensembles of LightGBM models with pooled training sets, quantile regression, and a mixed-integer linear programming reconciliation approach. Results highlight that spatiotemporal dependencies, as well as rounding bias, significantly affect forecast accuracy, underscoring the importance of integer forecasts for operational feasibility. Shapley analysis shows that short-term demand is reactive, shaped by life cycle maturity, autoregressive momentum, and operational signals, whereas medium-term demand reflects anticipatory drivers such as online engagement, planning targets, and competitive indicators, with online behavioral data considerably improving accuracy at disaggregated levels.

Figures (10)

• Figure 1: Summary plots for exploratory data analysis (Market-Product Line level).
• Figure 2: Summary plots for Market-Product Type level accuracy.
• Figure 3: Forecast accuracy of reconciliation methods across hierarchical levels.
• Figure 4: Distribution of SHAP values per feature for forecast steps 1 and 6.
• Figure EC.1: POOL-SEL-BP hyperparameter tuning.