Walmart-Store-Sales-Forecasting-using-Statistical-Models-in-R

GOAL

To create a forecast model using R that accurately predicts sales and demand for Walmart, incorporating factors like economic conditions and the effects of promotional markdowns and holidays.

KPIs:

Weekly Sales, Fuel Price, Unemployment, Week number, Event(categorical), Month(categorical)

SOLUTION

1. We gathered data, performed data exploration, made different datasets with required columns, plotted graphs, removed outliers, correlation analysis, made training and test datasets.
2. We planned on building 3 forecast models: Multilinear Regression model, Random Forest Model, and Convolutional Neural Network Model.

INSIGHTS:

1. Out of these models Random Forest model gave us an accuracy of 93% which was the best model.
2. The analysis revealed that holidays such as the Super Bowl, Thanksgiving, and Labor Day have a significantly positive impact on sales. Christmas-related sales reveal that customers tend to make purchases in the week leading up to Christmas and not on Christmas.
3. Month emerges as an important independent variable for sales prediction.
4. Sales are generally higher in the second semester of each year.
5. Fuel price and unemployment, when considered together, have a more significant impact on sales than the Consumer Price Index alone.