Figure 9From: Extracting complements and substitutes from sales data: a network perspectiveInternal structure of the substitute roles of onions (4) (left) and of apples (23) (right), with product nodes connected by both the complementarity scores \((W_{ij}^{(c)})\) (in black) and the substitutability scores \((W_{ij}^{(s)})\) (in orange), where the line thickness corresponds to how high the scores are, and products in the same L1 category are shown in the same colourBack to article page