How to find the insights from data
Business data analysis class started again at Yokohama National University.
The class is designed mainly for the 1st and 2nd year international students to give business data analysis (problem-solving) skills.
We focused on some basic techniques and thinking process to process raw data to get some insights from the data.
This was one of the exercises we used in the class:
The students were required to make their own “conclusion (not just findings nor analysis results) with Excel.
In this case, the output (goal) parameter is “frequency of visit”, while the input (explanatory) parameters are (i) Gender and (ii) Direct Mail (D/M) received or not.
In order to get a first insight from the data, you may compare the data on the each parameter (in this case, gender and D/M received). Let’s look at examples of the outputs:
In order to simplify the original data, averages are used in the upper matrix.
The following bar charts show the comparing results with the two parameters (gender and D/M).
Which parameter is more critical to the goal? Look at the significance of the difference!
(D/M received) has more gap between “YES” and “NO”, rather than one between the genders.
So you may focus on the D/M effects more than gender difference. (Let’s call it “Key parameter“)
Then you may dig into the D/M effects to find more detail (by splitting by gender).
This is an example of a conclusion drawn from the chart above:
” Direct Mail affects the frequency of the visits, while only female customers have such effect. Therefore, my proposal is to keep sending D/M to the potential female customers to get more visitors”
Again, here are the points to learn from the exercise:
- How to find the key parameter to dig into further?
=> Find the parameter with “larger gap” which indicates more critical factor to the goal/issue.
- What to do once you find the key parameter?
=>Split the data with another parameter to dig into the “Key parameter”