The world’s Largest Sharp Brain Virtual Experts Marketplace Just a click Away
Levels Tought:
Elementary,Middle School,High School,College,University,PHD
| Teaching Since: | Apr 2017 |
| Last Sign in: | 57 Weeks Ago, 2 Days Ago |
| Questions Answered: | 7570 |
| Tutorials Posted: | 7352 |
BS,MBA, PHD
Adelphi University/Devry
Apr-2000 - Mar-2005
HOD ,Professor
Adelphi University
Sep-2007 - Apr-2017
Telco Customer Churn Problem
“Churn” refers to customers switching to different providers. It is not a preferable situation to telecommunication service providers. Telecommunication companies are always interested in predicting which customer may churn, so hopefully they can take measures on time to reduce the possibility of churning.
The attached Excel file is a data set obtained from a telecommunication company. The data set contains data for 4708 customers. Variables include:
|
Variable Name |
Description |
|
account_id |
Customers’ account ID |
|
target_churn |
This is a binary variable. 1 means customer churned, and 0 means not. |
|
current_billamount |
A customer’s current monthly bill amount. Note these are business customers. |
|
Avg_Calls |
Average daily calls in minutes |
|
Avg_Calls_Weekdays |
Average daily calls in minutes for weekdays |
|
Percent_Increase_MOM |
Percentage of bill amount increase month by month |
|
Account_Age |
The time with the telecommunication company in the number of months |
|
acct_plan_type |
Gold or Silver |
|
Complaint_Code |
Multiple reasons why a customer may complain, e.g., call quality, billing problem. |
|
Current_TechSupComplaints |
The number of current technical support complaints |
|
Current_Days_OpenWorkOrders |
The number of days a work order is open |
|
Equipment_Age |
The time the customer has the equipment, in the number of months. |
|
Condition_of_Current_Handset |
1 – working, 0 – not working. |
|
Avg_Hours_WorkOrderOpenned |
Self explanatory |
|
Avg_Days_Delinquent |
Self explanatory |
Part One – What is going one?
You are a business analyst just graduated from the prestigious Business Analytics program at the Old Dominion University. Your boss, John, is counting on you to solve problems. John is a telecommunication veteran but he is neither a statistician nor a business analyst. He just passed this dataset to you and his first question is – can you tell me what is going on with these customers?
So, use appropriate descriptive statistics tools and techniques, generate a short but precise report for John so that he understands the overall picture. While John did not limit you to any specific software package, it is recommended that you use IBM SPSS Modeler. John is not expecting a 30-page report, but instead, he wants a couple of “slides.” For example, he may want to see one slide to tell him some key statistical measure, e.g., average monthly bill amount, percentage of customers churning. A combination of charts and numerical measures will be perfect. John is also interested in some new functionality IBM SPSS Modeler has, for example, a histogram with certain variables overlayed.
Part Two – Explore for possible reasons why customers churn
John is more interested in learning why customers churn. Of course, that is a difficult question. John thinks the first step is to explore the data for interesting patterns. Those patterns may spark some novel insights on why customers churn. John does not request a comprehensive report at this moment. For now, his questions are like: are customers with higher monthly bill amount more likely to churn or not? Is churning related to complaint code? Or maybe some factors combined together are related to churning?
So, part two of the report is to conduct some explorative study on the data set. You are not limited by John’s questions at all. Feel free to report any interesting findings to John. Again, you should use both charts and numerical measures to build your case.
Note – This is a time-limited assignment, you need to turn in your work before the deadline. There is no need to pay excessive attention to the format of your document. John’s suggestion is to use bullet points, and support the points with analyses and charts. For example, you may have a bullet point that says “Gold plan customers are more likely to churn than Silver plan customers” (note this is a made-up statement, not necessarily supported by data) and then support this point with a chart or some statistical analysis. Since this is not meant to be a comprehensive report, you need to be precise and concise. You can submit your report in either PowerPoint or Word document format. The length of your report should be 5~8 pages (or slides).
To complete this report, you are allowed to refer to textbooks or other course materials. You can also search the Internet, libraries, or other sources. But you are strictly prohibited from seeking help from others. Nor should you attempt to copy/paste, or paraphrasing materials from other sources. In other words, it has to be your own work.
Grading Standard
The standard for grading is very straightforward: the quality of the report. Quality is measured by a number of factors:
· Important information is identified and presented.
· Information is well organized. Readers can see the points, but not just fragmented information being thrown around.
· Information must be supported by appropriate analysis (including both numerical analysis and visual/graphical analysis). Making unsubstantiated statements will lead to point reduction. In other words, bullet points not supported by correct analysis will lead to a lower score.
· Proficiency in descriptive analytics is demonstrated.
o Readers can sense that the report writer is familiar with IBM SPSS Modeler. The report writer knows how to load a data file and specify data types. It is necessary to demonstrate familiarity with a variety of descriptive analytic tools and techniques. More importantly, the report writer should know when to use what tools and techniques.
Submission
Submit your work through Blackboard by the deadline. Note that the submission link will disappear after the deadline.
BNA-----------L 4-----------15 -----------Mid-----------ter-----------m E-----------xam----------- 1 -----------Tel-----------co -----------Cus-----------tom-----------er -----------Chu-----------rn -----------Pro-----------ble-----------m-----------