AN INNOVATIVE TIME SERIES BASED METHOD OF FORECASTING MONTHLY SALES OF CHAMPAGNE

Abstract

Time series analysis is a process which deals with data to identify trends and forecast future happening. The data used in this analysis is known time series data which is in a time interval format. The data was used is based on the data which is taken in a different interval of time, known as time-series data. Balancing of demand and supply is based on the accurate prediction of sales in future, if there is a lack of efficient forecasting than it can a challenge to run the business and make a good profit. Time-series forecasting is one of the most commonly used approaches in day-to-day in many organizations. Marketing research also uses time-series data to find their future predictions. Its main strength is to study the change of various event in time so that it can be a strong tool for marketers. Approaches of time series analysis are autoregressive integrated moving average(ARIMA), seasonal autoregressive integrated moving average(SARIMA), autoregressive moving average(ARMA), moving average(MA) and autoregression(AR). This paper consists of predicting the monthly sales of champagne by using time series and will also predict the future monthly sales of champagne. ARIMA and SARIMAX model was used for forecasting and predicted the sales of champagne for 10 years. Champagne dataset was used in this research and using time series, predicting model was prepared. It is predicting good and giving good results. Keyword: time series analysis, forecasting, sales prediction, future prediction An innovative time series based method of forecasting monthly sales of champagne Page No. 78 INTRODUCTION The main purpose of this paper is to find the future forecasting of the monthly sales of champagne. ARIMA and SARIMA model were used to make a model and which would be useful to predict the monthly sales of French champagne. We will find a plot which shows the prediction for The monthly sales of the champagne dataset is a time series dataset, which can be used to find the future prediction based on the past data points of the dataset. The dataset consists of months and total sales from 01/1964 to 09/1972. There are 104 entries, and the sales are counted in millions[8]. Forecasting is the process of determining the future happening. Forecasting can be used in many fields such as what is the GDP rate after five years, profits or loss after one year, decrease or increase in shares of a company, demand of a product after six months, climate forecasting. Forecasting is required to run an organization, so if there is an unfortunate happening in the future so it can be handled easily. The data which is recorded based on time is called time-series data. Examples are the annual income of an organization, the demand for a product, day-to-day temperature, and the number of passengers travelling. The data which is calculated per second, per minute, hourly, daily, weekly, monthly, yearly are all comes under time-series data. It can be used to do better planning and can suggest what should be done to increase profit and productivity. Taking a real-world example, if a shopkeeper knows the future sells of their products, so it is easy to calculate the amount of product to keep in stock. Owner of a website knows the number of users visit the website so it can be useful to handle website traffic efficiently. The model of time series predicts future values according to previous values. On the other hand, multivariate analysis is an analysis which involves finding one or more statistical outcome at a time. Time series are used in various fields such as in signal processing, econometrics, finance, astronomy, meteorology, communication, earthquake prediction and others. ARIMA models are the types of models which is used to forecast future prediction using time series data. The main purpose of this model is to describe and find the autocorrelations in the time series data. For example, in a company how many average products sold in a year Figure 1, we can see that the black line shows the actual product sold in years and the blue lines indicates the data which is predicted by the time series model, by this we can understand that we are using past data to predict the future sales. Aseen Saxena, Siddharth Nanda Page No. 79 Figure 1 : Average sales of product of a company