Another project created during Udemy's Tableau's course. I found this section particular interesting as it displays the analytical power held within tableau. The goal of this was to provide analytics on 10 new locations a company recently opened and determine which stores are the best option to invest more marketing into. The problem seems to be a standard analysis of return on investment question, and while straightforward scatterplots helped develop some insight into the problem, data mining can dig into deeper patterns within the data.
Cluster analysis was a perfect choice for finding a solution to this problem as it help separate the Revenue to Market Spending data into inherent clusters via the K-Means. City population data was joined to the original dataset to provide more data for K-means to work off, along with revenue and
Two views of the dashboard are located below. These contain a map of each of the company's regions and their summary statistics for vital information. Below this is the cluster analysis that Tableau found when utilizing revenue, market spend, and population as variables. To the right of these we have some cluster analysis details: A map of how the cities lay across the US, colored by specific cluster they fell within, and two charts which provide a look at what deterministic levels the clusters included for each variables. The second image of the dashboard filters the dashboard to display only information relating to the new locations in question. Linear regression's provides an easy answer to the initial business problem: The three stores that would benefit most from increased market spending would be the three located in cluster 2 (orange), as they are the ones with the largest weight per dollar spent in marketing (~7*market_spending).
Data used was provided from SuperDataScience.com/Tableau, under Section 8.