Archive for March, 2010

Application of Business Analytics

March 16, 2010

Data Mining

Data mining is the process of analyzing data and making it into useful information. This information can help to increase revenue or decrease cost. Technically, data mining is the process of finding correlations or patterns in large databases of different fields.

 Below is an example I’ve thought of to apply data mining techniques in the retail industry.

Business Understanding

OG Pte Ltd is a departmental store that sells a variety of items such as clothes for adults and children, cosmetics, homestyle, travel goods, bags etc. 

The management of OG Pte Ltd realizes the section selling bags have the lowest sales volume.  As such, promotions have to be executed to increase the sales volume. In order to achieve successful promotions, there is a need to find out the target group, the different categories of customers who purchase the bags, who are those who don’t purchase the bags, demographics etc.

Data Preparation

The data required for analysis purposes can be collected in different ways such as point-of-sale records, surveys, membership applications and transaction volume. It is important to know the types of data required for the analysis.

 Variables

The following is a list of variables that’s necessary for identifying target groups.

 Target variable:

  • Purchase bag –Yes or No

Predictor variables:

  • ID – For data count
  • Gender – male or female
  •  Monthly Income – eg. 1) Less than $2000 2) $2000 to $5000 3) Above $5000
  • Age
  • Marital Status
  • Ethnicity
  • Occupation
  • Work location
  • OG member
  • Brand conscious – Yes or No 

  Modeling

Clustering

Modeling is required to perform data analysis. For analysis purposes, clustering can be use to group similar objects into the same cluster and dissimilar objects into different clusters. It is useful for market segmentation and aids in executing promotions on target groups.

 By using clustering, it helps to understand the purchasing patterns of the customers, behavior and needs of customers segments. Clustering helps us to find out things like, which cluster is more likely to purchase the bags. One of the clustering method is K-Means.

 Decision Tree

However it is not enough to use clustering alone because it can’t tell us which variables are more important. With decision tree, it can tell us things like, is age or income a more important variable to identify customers purchase a bag. This can be interpreted by looking at the decision tree, the higher the input variable is in the tree, the more important it is. Decision tree is also easier to interpret compared to other models like neural network. Decision Tree can generate classification and regression trees.  Classification is generated when target variable is categorical. Regression tree is generated when target variable is continuous.

 Conclusion

Base on the result of clustering and decision tree, OG Pte Ltd is able to come out with promotions that attract the particular group of customers and satisfy their needs.

Advertisements

Hello world!

March 7, 2010

Welcome to WordPress.com. This is your first post. Edit or delete it and start blogging!