Subscribe to DSC Newsletter

Better Banking with help of Analytics and Machine learning

In 2015, I was working at Diebold where we build ATM machine hardware and software and complete ecosystem around the ATM. When we talk about ATM machine, it is a collection of very complex small hardware which collectively performs tasks. And typically, when we think ATM is only used for cash withdraw and that is not true. When we talk about ATM it is a Bank branch itself. You can deposit cash, withdraw cash, deposit cheques. And Whatever we can do in the branch we can do with ATM.

When we were developing an application in mid-2015, which runs on ATM Machine stack I came up with an idea of predictive maintenance of ATM Machine.

There is already a solution for maintenance of ATM, but they are not predictive. Once any hardware fault happens on the machine or Machine goes offline a ticket is generated and then it is up to the team to visit the ATM and fix the issue. This leads to the machine being offline or not working or partial working mode for at least couple of hours.

This couple of hours turn into days when we talk about remote places/rural areas/out of city location in India. and we all know ATM and bank presence in the remote area is very less.

when we visit sometime in rural INDIA I have seen long queues in some ATM and while some ATM is out of order due to hardware or out of money. This is one of the big issues that why most of the rural INDIA does not want to depend on Bank and putting cash in the bank.Most of ATM fails working due to improper usage for that location or may be weather condition such as dust or electric fluctuations.

Now coming back to one of solution we can think of. There is different hardware component in ATM which generate logs and there is a system installed on ATM which pulls that logs to the server and keeps that log for later use or for compliance purpose.

This is Tera-Byte data which we can try to upload into Hadoop cluster and run predictive maintenance Machine Learning algorithm and text analytics algorithm to predict which ATM hardware is going to fail.

There are additional data such as the old history of hardware failures/replacement/ink refill/ paper changes etc. which can help us to predict more accurately.

And apart from above-mentioned data, we can add hardware’s manufacture date and their life time.

we can predict and due to prediction bank can use their resource very efficiently. Following are a scenario where this model and data can be useful.

1. health monitoring of ATM Network

2. Predict Hardware failure

3. Find out Trend of ATM usage

4.ATM response Time

5. Hardware life time

5. Prediction of ATM Failure

6. Refill of ink/ refill of paper etc.

 

The second solution for this using IOT in ATM Machine and consider the scenario when ATM printer detects that it is going to out of paper or ink and report to the server for a refill it or change it.

Or ATM Cash cassettes can notify server for less cash. This is also done currently using different way using maintain cash in ATM and withdrawal data.

 

Views: 2378

Comment

You need to be a member of Data Science Central to add comments!

Join Data Science Central

Videos

  • Add Videos
  • View All

© 2019   Data Science Central ®   Powered by

Badges  |  Report an Issue  |  Privacy Policy  |  Terms of Service