7. Data Models

Simple Linear Regression Model in Marketing Analysis

• Simple linear regression (SLR) is used to evaluate the relationship between independent and dependent variables in marketing analysis.
• The SLR model requires at least one quantitative independent variable and one quantitative dependent variable.
• The data for SLR must meet five assumptions: minimum sample size of 20, linearity, homogeneity of variance, normality, and independence.
• These assumptions ensure the accuracy of the results obtained from the SLR analysis.
• Before employing SLR in marketing analysis, it is crucial to verify that all the variable requirements and data assumptions are met.
• The basic steps for running SLR include creating a scatter plot for the independent and dependent variables and drawing a trend line to represent the data.
• In Tableau, a popular data visualization tool, the scatter plot and trend line can be created by dragging the variables to the appropriate shelves and selecting the "Trend Lines" option.
• Once the data is plotted and the trend line is drawn, predictions can be made using the trend line.
• An example visualization of a SLR conducted using Tableau showed the prediction that 60 Adware Clicks will yield around 6 Adware Conversions.
• The SLR model can be a valuable tool for making predictions in marketing analysis, particularly when evaluating the relationship between ad clicks and conversions.
• This type of analysis can provide insights into the effectiveness of marketing strategies and guide decision-making for optimizing ad campaigns.
• By following the characteristic purpose, variable requirements, and data assumptions, marketers can ensure the suitability and accuracy of employing SLR in their marketing analysis.

Key Points on K-means Cluster Analysis

• K-means clustering is the default method in various software programs, including Tableau, and involves calculating specific points within data to create groups by minimizing the distance from them.
• For cluster analysis, the data must include at least one quantitative independent variable and one quantitative dependent variable.
• The data assumptions for K-means cluster analysis include a minimum sample size of 50 data points per grouping, sphericity, homogeneity of variance, and equal prior probability.
• Sphericity in K-means clustering means that the groupings fall into a rounded area when plotted, forming rounded clusters around central data points.
• Homogeneity of variance assumption in K-means clustering implies roughly the same amount of variance in independent and dependent variables.
• Equal prior probability assumption in K-means clustering means each grouping should have roughly the same likelihood of occurring within the data.
• These assumptions serve as a checklist to determine the suitability of K-means cluster analysis for data analysis.
• Once determined suitable, the basic steps for running a K-means cluster analysis involve creating a scatter plot for the independent and dependent variables and then clustering the data into the desired groups.
• In Tableau, the scatter plot for K-means cluster analysis is created by placing the independent variable on the X-axis and the dependent variable on the Y-axis, followed by clustering the data through the "Analytics" tab.
• A sample visualization for a cluster analysis using Tableau involves analyzing customer spending data to define groups based on spending, as shown in the scatter plot.
• An interesting observation from the analysis is that the group of lowest spenders also tends to have the highest number of clicks.
• The text provides practical guidance and examples for conducting K-means cluster analysis using Tableau, emphasizing its application in marketing analysis.