The aim of this project is to leverage data and suggest "more promising" prices for Kellogg's cereals that would not only solidify Kellog's market dominance but also lead to increased profits.
Specifically, the project consists of the following parts:
-
Visualizations: We explore the data using visualization tools to study the pricing behaviors for the Kellog's cereals
-
Optimal Pricing Setting: We conduct regression analyses to come up with optimal prices for Family cereal and Kids cereal respectively
-
Competition Analysis: We conduct regression analyses to estimate bias induced by not considering competition
-
Predictive Modeling: We construct a sales prediction model where one of the explanatory variables is price. Using the optimal prices we come up with in the previous step, we predict the sales and calculate estimated profits. Comparing this to the profits obtained by the current prices, we verify the competency of the new prices.
Below are the folders and files created for this project:
-
Data: This folder contains original data we obtained from dunnhumby as well as the data we transformed for the analyses
- dunnhumby _Breakfast-at-the-Frat: original data (i.e., Store, Products, Transaction datasets) and descriptions
- transaction_cereal.csv: subseted from Transaction dataset to only contain data for cereals
- cereal_merged_df.csv: merged dataset of Store, Products, and subsetted Transaction
- cereal_Kellogg_Month.csv: populated time indicator columns to use in visualizations
-
Visualization: This folder contains scripts that perform exploratory data analysis and visualizations
-
price.ipynb: This script performs regression analysis to come up with the optimal prices and to estimate Bias induced by not considering competition
-
project_modeling.ipynb: This script constructs sales prediction models, calculates estimated profits, and performs comparison on profits generated by current prices and new prices
For the overall summarized results, please see presentation.pdf.