The problems provides us with a dataset with 6700 observations and 17 variables. We need to decide which customer is going to revert positively on the new product launched by the company. Details of the user are provided on the basis of their history with the banks.
The data is highly messy and missing. Extensive amount of time and efforts went into data cleaning and wrangling before it was ready to apply ML algorithms.
It can be seen that around 17 clusters can be formed with minimum errors for this data.