Project: Kalbe X Rakamin Data Scientist

Problem Statement

Inventory Team:

Predicting the sales quantity (quantity) for the overall Kalbe products.

• Knowing the estimated quantity of products sold, the inventory team can create sufficient daily stock supplies.
• The prediction should be done on a daily basis.

  Marketing Team:

  Creating clusters/segments of customers based on several criteria.

• Creating customer segments.
• The marketing team will use these customer segments to provide personalized promotions and sales treatment.

Tools:

• Python
• Jupyter Notebook
• Tableau
• Dbeaver
• PostgreSQL

Exploration with PostgreSQL and Dbeaver

Average Customer Age based on Marital Status:

• Married = 43
• Single = 29

Average Customer Age based on Gender:

• Woman = 40
• Man = 39

Sotre with Highest Total Quantity: Lingga

Top Product by Total Amount: Cheese Stick

Tableau Dashboard

For interactive use, please visit: Tableau

Machine Learning

Time Series Model

MAE is found to be 0.24, meaning the model’s forecasts differ from the actual values by approximately 0.24 units. MAPE is 6.53%, meaning the model’s forecasts deviate from the actual values by about 6.53%. MSE is 0.08, meaning the model’s forecasts have, on average, squared differences of 0.08 with the actual values. RMSE is 0.29, meaning the model’s forecasts have a root mean squared difference of approximately 0.29 with the actual values.

Clustering Model

From the 2-cluster division, it is found that in cluster 1, customers have a higher preference for Thai Tea, Ginger Candy, and Oat. Meanwhile, in cluster 0, customers have a higher preference for Cheese Stick, Choco bar, Crackers, and Cashew.

Code source: Github