Sales data contains information about completed sales, not about missed sales. Or at least?

Vrumona is one of the largest suppliers of soft drinks, waters and juices in the Netherlands, and supplies their products to all major supermarket chains, including Albert Heijn. Based on a dynamic K-nearest neighbors model, we have created a model in Python that calculates for each Albert Heijn whether and how much improvement in sales per product group is still possible. To determine the right features, we first investigated what the sales of a specific soft drink or soft drink packaging correlates with. To give an example: in the different areas of the Netherlands, people appear to have different preferences for soft drinks. Rivella, for example, is mainly drunk in the north. In addition, relatively more cans are bought in large cities than in villages.
With our model, Vrumona can calculate how big the missed opportunities are. The results are displayed in clear visualizations in a Power BI report per week by account manager, so that action can be taken to turn missed sales into maximized opportunities.
