Homework 13
Question 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
...
Homework 13
Question 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?
Objective:
The objectives of the algorithms are
1). to identify who should be shutoff, and who should not,
2). and from the groups that should be shutoff, how to group them together in a way that it’s most effective from
scheduling perspective.
Important data:
• Income
• Household size
• Monthly utility amount
• Payment history (last 24months)
• Location
• # of past due - whether this customer has a history of late payments
• Past due amount – how much was the past due amount
• # of months with past due – in the past, how many months before this customer catch up with their
payment again
• Is Autopay setup – Sometimes it could be due to autopay failure due to credit card information changes, etc.
• Method of payment in the past - This would help to see if there’s any change in their payment ability (i.e.
switch from check to credit card)
• Contacted by customer service – have we tried to contact them prior?
• # of months take for reactivate in the past – if they had delinquency before, how long did they catch up
• Travel cost
• Labor cost
Model 1: Determine who should shutoff
GIVEN These Data:
• Payment history
• Is Autopay setup
• Past due amount
• Monthly Utility amount
• # of months with past due
• Method of payment in the past
• Contacted by Customer Service
USE:
• Holt-Winters
TO:
• I will use the triple exponential smoothing (Holt-Winters) algorithm TO determine if they can afford to pay
for the utilities, since this algorithm will take into account of both seasonality, and trends, especially taking
the recent months payment history with greater effect. If they had been beyond their typical late payment
past due months, then they would be more likely to miss because they might be in a more difficult financial
situation, especially if their monthly bill amount is high.
Model 2: Grouping the shutoffs
GIVEN These Data:
• # of months take for reactivate in the past
• # of months with past due
• Location
• Travel Cost
• Labor Cost
• Past due amount – we might want to shut them down sooner, even they have very high balance amount
USE:
• Linear regression and K-Nearest Neighbor
TO:
• I will use the Linear regression algorithm TO first determine the probabilities of total repayment effort. And
then use KNN to use this repayment effort along with the other attributes to create the clusters based on
the sum of the total financial cost ( lost of payment, travel cost, labor cost).
[Show More]