About the Customer
Industry: Building Materials
The client, a pioneer of ceramic tiles in India, is an important division of one of India’s leading integrated building materials companies, with a variety of products from cement, concrete, tiles, bathroom products to modular kitchens. It enjoys the reputation of being the only entity in India to offer end-to-end lifestyle solutions.
The major challenge that the client encountered was to truly predict, with facts and figures, the sales for the next year even if there was ample live data available for all the previous years. There was an immediate need for a tool which could correctly predict the futuristic figures and give the management an insight into the further opportunities.
The main objective of the solution provided by SPEC INDIA was to develop and implement a Predictive model which can predict the company sales and city wise sales for the next year, based on the past data of the organization stored in the live dataset. This solution involves many data mining, predictive modeling and analytic techniques to bring together the management, IT and modeling business process to make predictions about future.
Key Process Areas
- Project Definition
- Define project outcome – predict next year sales for company
- Scope – provide better understanding for future decision making
- Business objective – extract profit and targets of next year’s sales
- Identify data set – identify parameters affected by the sales
- Data Collection
- Collect company data from multiple sources for analysis to predict sales for coming year
- Collect most affected parameters like sales year, dealer code, dealer information, employee code, employee visit, sale amount, target of each sale, time spent on each visit etc.
- Data Analysis
- Collection of useful information from different tables and cleansing of data like missing values, inappropriate data, formatting data, thereby extracting a useful data set for the prediction model.
- Statistical Analysis
- Create a statistical model to predict the next year’s sale, with multiple linear regression algorithm
- Validate the assumption, hypothesis and test it using standard statistical models
- Data Modeling
- Create accurate predictive model predicting the future sales amount of the company and yet another creating city wise prediction analysis
- Validate the accuracy of these models through MLR validation techniques
- Deploy the analytical results of the model for company to easily visualize and identify next year’s company sales, through a predictive dashboard
Tools and Technologies
SPEC INDIA’s successful implementation of Predictive Analytics Model helped company in easily and accurately forecasting what the next year sales could be, with effective facts and figures to support it. It also included city wise sales forecasting. This is a sure shot advantage to the management to delve deeper into the predicted information and take business decisions accordingly.