The case study is to come up with project implementation methodology with forecasting models which gives best demand forecast using the historical training and validating on the hold-out set.
Our objective is to predict demand for a FMCG company, taking into consideration the influence of promotional schemes provided by Distributors/retailers. This entails solving the issue of demand planning and promotion planning simultaneously.
As of now, Team-Datacurate is exclusively composed of its employees so we cannot recruit individuals from external sources to join the team.
How can we incorporate into the model promotion/schemes data, which is very diverse, and for some SKU’s unique as well i.e. occurring only once, making it difficult for us to categorize them? Is there any method to club such promotion data into fewer categories?
