Homework 12
Question 18.1
Describe analytics models and data that could be used to make good recommendations to the power
company.
Here are some questions to consider:
• The bottom-line question is which shutoffs sh
...
Homework 12
Question 18.1
Describe analytics models and data that could be used to make good recommendations to the power
company.
Here are some questions to consider:
• The bottom-line question is which shutoffs should be done each month, given the capacity
constraints. One consideration is that some of the capacity – the workers’ time – is taken up
by travel, so maybe the shutoffs can be scheduled in a way that increases the number of them
that can be done.
• Not every shutoff is equal. Some shutoffs shouldn’t be done at all, because if the power is left
on, those people are likely to pay the bill eventually. How can you identify which shutoffs
should or shouldn’t be done? And among the ones to shut off, how should they be prioritized?
The main goal of this analysis is that we want to find which shutoffs should be done each month.
To define which shutoff should be done or not, we need to find or to predict which customers
will not pay. Thus, the first step in my approach will be to analyze the customer base of the
company and try to define different groups of clients:
- A group of regular and good payers
- A group of non-regular payers
I) Find the delinquent customers
For this first classification, I would use logistic regression.
Given:
- Credit score
- Income
- Historic default on payment (all power companies)
- Average time of bill payment (- x days before the due date or + x days after the due
date)
- Average monthly bill
Use: Support Vector Machine
To: Define the classification for each customer into 2 groups: “will pay” and “will not pay” the
bills.
Now, I will focus on the “will not pay” group defined.
This study source was downloaded by 100000834091502 from CourseHero.com on 05-16-2022 06:47:13 GMT -05:00
https://www.coursehero.com/file/57498559/HW12power-companypdf/
In this group, there are also different sub-groups of potential non-payers with different profiles.
And regarding these different profiles, the target for shutoff will not be the same because some
of them will eventually pay:
Profiles of non-payers Perspectives Target for shutoff
Customers with variable
incomes (seasonality)
Eventually will pay late No
Disorganized customers
(forget everything!)
Eventually will pay late No
Customers with financial
inability to pay
Will not pay No, because of the special aid
program
Customers with financial
ability to pay but don’t pay
Will not pay Yes
Then, on this group of customers, I want to extract the “will not pay” customers with Target
“Yes”. Assuming there is a special program to help people with financial inability, I’ll be able
to extract this special group using the criteria required to be in the program and set them aside.
Then, on the rest of the group and based on the idea that some of them will pay after a while:
Given:
- Credit score
- Income
- Historic default on payment (all power companies)
- Average time of bill payment (- x days before the due date or + x days after the due
date)
- Average monthly bill
Use: Logistic Regression
To: Define the probability that a customer will pay after a while (for example using the average
time of bill payment when the due date
[Show More]
.png)
