D-mart is one of the largest retailers in India and it is very important for them to have accurate forecasts for their sales in various departments. Since there can be many factors that can affect the sales for every department, it becomes imperative that we identify the key factors that play a part in driving the sales and use them to develop a model that can help in forecasting the sales with some accuracy.
- This prediction based on sales of past 10 years annual sales data.
- Model is able to predict the sales of the coming year with an accuracy of approx. 79%, which turns out to be a reliable figure.
- Model require inputs like: - Item’s weight, type, fat content, identifier code, MRP, Outlet’s size, type, location tier and identifier code to calculate the sales accurately.
Model creation :-
- Python
- Scikit learn
- Python libraries like – pandas, numpy and matplotlib
UI :-
- Front end – HTML and CSS
Frameork
- Django
- Data is downloaded from Kaggle and cleaned for further analysis.
- Using correlation function, attributes required for prediction are sorted
- Label encoder and one hot encoding is used on the attributes for further processing
- Linear regression and Multiple regression model is created which generates an accuracy of about 79%