Sam works as a manager in a national service center of a software company. The company sells packaged design software. Customers who buy the product require service in case they face any problem in the software. Sam is concerned about the capacity of the national service center to handle to number of service requests. There are two types of service requests, one customer submits written request which is called as STRs and two they can call national service center which are denoted as SCALLs. Sam is concerned about SCALLs as they have to handled by service personnel on interactive basis. The number of calls received per month varies and Sam wants to predict number of calls based on number of software packages sold. This will help Sam and his department to arrange for required number of service personnel for this and other products. Once the procedure for this product is established it can be replicated to other products.
There are two ways to look at this problem.
1. Predict the number of service calls i.e. SCALLs based on the number of packages sold in the month under consideration. 2. Create a time lag model based on the fact that a customer may face a problem in the month subsequent to the month(s) in which he/she purchased the software.
No Lag model:
Let us look at first model. Here we try to predict number of service calls based on the number of packages sold in the respective month. The scatter plot and histogram (see the excel sheet) as follows shows the number of product sold against the number of service calls received
From the above two figures it is difficult make any comment on the relationship between number of packages sold and SCALLs. Hence, we correlate the two variables. The correlation results are as follows
|Correlation analysis | | | |Q |SCALL | |Q |1 |...