 |
| #Chaid a Skillset for Data Science Engineers: |
CHAID analysis is especially useful for data expressing categorized values instead of continuous values. For this kind of data some common statistical tools such as regression are not applicable and CHAID analysis is a perfect tool to discover the relationship between variables. One of the outstanding advantages of CHAID analysis is that it can visualize the relationship between the target (dependent) variable and the related factors with a tree
image.
Different Scenarios where CHAID analysis can be used:
CHAID Analysis for Surveys -Most survey answers have categorized values instead of continuous values. Finding out the statistical relationship in this kind of data is a challenge. CHAID Analysis is a good solution for such survey studies.
Related: How to Make Career Change from Mainframe to Data Analyst
CHAID Analysis for Customer Profiling -Based on historical customer data, CHAID Analysis can be used to analyze all characteristics within the file, for example, product/service purchased, dollar amount spent, major demographics and demography of the customers, and so on.
A blue print can be produced to provide an understanding of the customer profile: strong or weak sales of products/services; active or in-active customers; factors affecting customers’ decisions or preferences, and so on. Such customer profile will give the Sales & Marketing Team a clear picture of which type of person is most likely to buy the products and services based on factual purchase history, geo-demographics and lifestyle attributes.
CHAID Analysis for Customer Targeting -Recruiting new customers via direct contact (phone or mail) is a time consuming and costly effort. For most products or services, the hit rate is less than 1%. That means, in order to get a new customer, over one hundred contacts are required.
By mapping the current customer list to a general population database (e.g., SMR Residential Database that contains 12 million listed households), CHAID Analysis can find the household clusters that have much higher incidence rates than the average. By concentrating on these household clusters, the actual hit rate can be dramatically raised. The result is “Fewer phone calls or mail pieces with higher sales returns!” .
Related:
No comments:
Post a Comment
Thanks for your message. We will get back you.