MySQL

• we solve the My SQL tasks
• pass two taks to solve it later

Python

• Step 0
• we read the file into Python
• Create the column names because the csv has no headers

Step 1

• Explore the data (with Python,Tableau,Excel)
• Missing Values: We drop the Missing values : only 24 values yes and no are the same missbalanced in comparison to the whole dataset
• There are no duplicates in the Dataset
• unwanted columns Q1_Balance to Q4_Balance and customer_number
• target variable: offered accepted
• Maybe later: create groups
• datatypes are correct

Step3 EDA

y-Variable(yes, No)

• one hint: Drop the outliers for NO can redurce the dataset. The dataset für ysy gets notr bigger but the relationship between no and yes gets better. It is a step to imporve the dataset befor starts the model.
• we create 3 different dataset for handling with the outliers and we how our model will works (because we have different housholdsize and differentz income level that influence of course the dataset

• we will check if the houeholdsize 8 and 9 (only 1 row) stays in the dataset. Because if we normalize/encode the features, the influence of the bothe rows withe the householdsize 8 and 9 have a big influence of the new interval.

• Result of handling the outliers and household_size: The household_size is not a outliers. This amount stays in the datset after reove the outliers. Additionally to that, it is easier to remove the outliers from the columsn q1_balance to q4_balance that to remove the outliers only from the column average_column. So we have to drop the single row for houehold_size8.