Predictive Analysis of Fabric Softener Brand
Datasource: Dataset Source: Research Paper of Fader, Peter S. and Bruce G.S. Hardie (1996), "Modeling Consumer Choice Among SKUs," Journal of Marketing Research, 33 (November), 442-452. (Dataset: 9781 Records, 21 Predictors)
The data come from an IRI panel in Philadelphia and cover the period from January 1990 to June 1992. The only criterion for inclusion in our data set is that the household must have made at least one purchase in 1991. This gives us 594 qualifying households that made a total of 9781 purchases over the 2.5-year period. The 3227 purchases from 1990 are used for initializing model variables, and the 4417 purchases from 1991 are used for model calibration. The remaining 2137 purchases made during the first half of 1992 are used for forecasting purposes.
Tools used: R studio, Excel 2013
Skills: R, Statistics – Multinomial Logit & MA forecasting, Tableau.
Tasks Achieved: • Cleaning and preparing the data to run the model. Some dummy variables assigned. • Build multinomial logistic model to perform brand analysis and to forecast customer preferences based on SKU • Analyze customer loyalty and Sales dependency based on SKU attributes and price variables. • Forecast Sales using moving average, interpret customer behavior and sales trend based on brands and SKU.