ikea_datathon_2023
The present repository shows our winner proposal for the Datathon organized by IKEA, which consisted of two main challenges:
1. Waste Challenge
Understanding, predicting and reducing food waste by IKEA's restaurants, in line with the Swedish company's commitment to sustainability, cost consciousness and leading by example.
Winner idea: Transform the amount of wasted food by category to the amount of money wasted by each category, in order to create a dashboard aligned with the mentioned objectives.
2. Visit Challenge
Forecasting the amount of visits per day of the month of March 2023 in the Barcelona and Madrid stores, as a manner to optimize the resources needed in terms of inventory, employees and even to improve the Marketing strategies.
Winner idea: Substitute the empty values of the dataset with the average of its correspondent day of the week (as the stationality is weekly) and optimize the hyperparameters of an autoregresor based on the Gradient Boosting algorithm.