🍐 我们总结了美国代写中——NYU代写Business Analytics的经典案例，如果你有任何代写project的需要，可以随时联络我们。CoursePear™ From @2009。
Business Analytics, MG-GY 8413, Fall 2022
What is the CFPB?
The Consumer Financial Protection Bureau (CFPB) is a U.S. government agency that makes sure banks, lenders, and other financial companies treat their customers fairly. The CFPB was created in 2010 by the Dodd Frank Act in the aftermath of the financial crisis.
One of the primary functions of the Bureau of Consumer Financial Protection (CFPB) is collecting, investigating, and responding to consumer complaints.
When consumers submit complaints online or over the phone, the CFPB asks them to identify the consumer financial product or service with which they have a problem, as well as the type of problem they are having with that product or service. This provides information that the CFPB can use to analyze complaints.
The CFPB routes consumers’ complaints about financial products and services—and any documents they provide—directly to financial companies, and works to get consumers a timely response, generally within 15 days.
Consumers may then choose to dispute the resolution offered by the company.
The CFPB publishes the consumer’s narrative description of his or her experience if the consumer opts to share it publicly and after the CFPB takes steps to remove personal information. This data is available on the CFPB’s website.
The data is in the public domain, has over 2 million anonymized recent records, and covers 6000+ financial providers of all varieties. It has a number of features, including a text narrative of the complaint. The enitre dataset can be downloaded following the instructions at https://www.consumerfinance.gov/data-research/consumer-complaints/. The website also provides additional information on the data.
For this project, we will use only the data till 2017, and only for the top 5 banks in the US. In order to make sure we are all working off the same data, I have downloaded and filtered the data , and made it available on Jupyterhub. You do not need to download the data yourself. We will use the file complaints_25Nov21.csv available in Jupyterhub under the shared/ folder.
The task – A report to the “Big Banks’ Association”*1
The five largest banks in the United States have come together as the “Big Banks’ Association” and engaged your consulting firm to provide them a report based on complaints data.
Analyze the complaints data for the top five banks in the US included in the file. You are required to create a report, no longer than 10 pages, covering the topics described below.
- A description of trends, key insights from the data such as problematic products, customers, dispute rates, types of issues etc. You are free to use all analytic and presentation tools available to you to produce a meaningful report.
- You will observe that in the majority of cases customers are satisfied by the explanation provided by the banks, and the resolution offered. In some cases, customers dispute the explanations provided (flagged in the ‘Consumer disputed?’ field). In these situations, the bank has to perform additional investigations, and possibly offer further relief to the customers. As a result, the cost of dealing with disputes can be quite high, without including potential reputational damage and the risk of regulatory action. On average, it costs the banks $100 to resolve, respond to and close a complaint that is not disputed2. On the other hand, it costs banks an extra $500 to resolve a complaint if it has been disputed. So the banks have also asked you to create a model that can help them identify complaints that will end in a dispute. Create and evaluate a predictive model that can help the banks identify future disputes so they can perform “extra diligence” during the first round of addressing the complaint with a view to avoiding eventual disputes. If a complaint is picked up for “extra diligence”, it costs the bank an extra $90 to perform the additional diligence steps. But if they correctly spot a future dispute in a timely way, they can avoid spending the extra $600 later to resolve the dispute. There are also non-quantifiable benefits from happier customers, and happier regulators. Hint: Think about Calculating Total Cost in Dollars
- The moment a complaint enters the CFPB’s system, there is $100 cost to resolve it. This applies to every complaint.
- After that, if a complaint’s resolution is disputed by the customer, an additional $500 has to be spent (for a total cost for such cases to be $600).
- But the bank can intervene in advance by spending an extra $90 for extra diligence, and that can make sure the complaint’s resolution is not disputed. But obviously, the bank would not want to spend this extra money on complaints that would not have been disputed anyway, Create a predictive model that can help the banks keep their complaint related costs low. You are free to use any tools or modeling techniques you believe would work for the banks. Present 1 * There is no such thing as the “Big Banks’ Association”. This figment of imagination helps us keep the problem tractable for analysis, and focus on a single type of company, which is the large US retail bank.
2 Cost estimates for dealing with complaints are hypothetical. These are intended to encourage students to think about the cost of false positives, and balance it with the costs of false negatives.
the results of your modeling in your report and how the model should be used by the banks, and how it would benefit them.
This report is to be addressed to the President of the Big Banks’ Association, New York. You can assume that this report will be shared with and read by the Executive Management3 of the 5 members of the Big Banks’ Association. A cover page, table of contents and the appendix do not count towards the 10 page limit. Be aware though that the executive managers may not have the time or patience to read through appendices, so be sure to highlight all you need to say in the body of your report! (If you are using Python, please submit your Jupyter notebook as a reference for me and Dongling, the class TA.)
What should the report contain?
You have complete editorial freedom to create the report in a way you think would be effective. I would be looking for readability, completeness of analysis, and elements such as:
- Is there an Executive Summary?
- Is the problem to be solved described accurately and succinctly?
- Is the data exploration reasonably comprehensive?
- Was a predictive model created?
- Was the difference between precision and recall recognized?
- Was the threshold tweaked/ROC curve reviewed to improve recall?
- Were “total costs” (or savings) calculated in US Dollars, accounting for both FP & FNs?
- Was a firm recommendation provided to management? The above is just guidance to get your creativity going, not a limitation on what you can do.
3 Executive Management represents the highest leadership levels in the company, such as the CEO and their directs. Executive Management generally reports to the Board of Directors.
Reproduced from https://www.consumerfinance.gov/complaint/data-use/
Date received Product
Consumer complaint narrative
Company public response
The date the CFPB received the complaint. For example, “05/25/2013.”
The type of product the consumer identified in the complaint. For example, “Checking or savings account” or “Student loan.”
The type of sub-product the consumer identified in the complaint. For example, “Checking account” or “Private student loan.”
The issue the consumer identified in the complaint. For example, “Managing an account” or “Struggling to repay your loan.”
The sub-issue the consumer identified in the complaint. For example, “Deposits and withdrawals” or “Problem lowering your monthly payments.”
Consumer complaint narrative is the consumer-submitted description of “what happened” from the complaint. Consumers must opt-in to share their narrative. We will not publish the narrative unless the consumer consents, and consumers can opt-out at any time. The CFPB takes reasonable steps to scrub personal information from each complaint that could be used to identify the consumer.
The company’s optional, public-facing response to a consumer’s complaint. Companies can choose to select a response from a pre-set list of options that will be posted on the public database. For example, “Company believes complaint is the result of an isolated error.”
The complaint is about this company. For example, “ABC Bank.”
The state of the mailing address provided by the consumer.
The mailing ZIP code provided by the consumer. This field may: i) include the first five digits of a ZIP code; ii) include the first three digits of a ZIP code (if the consumer consented to
Consumer consent provided?
Date sent to company
Company response to consumer
publication of their complaint narrative); or iii) be blank (if ZIP codes have been submitted with non-numeric values, if there are less than 20,000 people in a given ZIP code, or if the complaint has an address outside of the United States).
Data that supports easier searching and sorting of complaints submitted by or on behalf of consumers.
For example, complaints where the submitter reports the age of the consumer as 62 years or older are tagged “Older American.” Complaints submitted by or on behalf of a servicemember or the spouse or dependent of a servicemember are tagged “Servicemember.” Servicemember includes anyone who is active duty, National Guard, or Reservist, as well as anyone who previously served and is a veteran or retiree.
Identifies whether the consumer opted in to publish their complaint narrative. We do not publish the narrative unless the consumer consents, and consumers can opt-out at any time.
How the complaint was submitted to the CFPB. For example, “Web” or “Phone.”
The date the CFPB sent the complaint to the company.
This is how the company responded. For example, “Closed with explanation.”
Whether the company gave a timely response. For example, “Yes” or “No.”
Whether the consumer disputed the company’s response. The unique identification number for a complaint.