Running head: FINAL PROJECT MILESTONE THREE 1 Andrew Fontenot December 9, 2018 DAT 220: Fundamentals of Data Mining Final Project Milestone Three
FINAL PROJECT MILESTONE THREE 2 Analysis Organization Cluster analysis was used to segment the customer population into distinct groups of customers with similar attributes. This was done for characteristics such as age, income, marital status, and spend rate. This technique can be used to segment customers for purposes of developing ad campaigns for specific customers which may be more effective than targeting the entire population. Distinct customer segments were found using all the available characteristics. Customer groups formed by segmenting the population for amount spent at a restaurant showed that the more a customer spends at a restaurant, the more they tend to spend online when they make a web purchase. These results can be used target specific customer groups which have a higher return potential on marketing dollars. One limitation to this analysis is that I used only used a single variable to segment customers. Other cluster analysis techniques which use more than one variable may more effectively segment customers. Linear regression was used to predict web store spend. If a useful model can be developed to predict spend rate at any of the store types, the model can then be used to evaluate how different customer characteristics affect the amount spent. With this knowledge, Bubba Gump can then target specific customers which will spend the most or research how they can alter their stores to better appeal to customer which do not spend large amounts. I began by using all available continuous variables to predict web store spend but found that the only variables which were useful were restaurant spend amount and web site visits. With only these variables in
You've reached the end of your free preview.
Want to read all 5 pages?
- Summer '16
- Regression Analysis