Fig. 7: PDPs of key variables.
From: A causal discovery and inference framework for on-demand food delivery delays
The figure visualizes the marginal effects of the top causal variables on delivery delay using PDPs. In each panel, the PDP curve summarizes the model-averaged relationship between the focal variable and predicted delay while averaging over the distribution of other features; SHAP points are overlaid to reflect the distribution and heterogeneity of effects across observations. Interpretation focuses on the central data range (1st–99th percentile) to avoid extrapolation from extreme values.
