The factory manager, April Knapp, has stormed into Barry Sleepy’s office and tossed her resignation on the desk. “I’ve had it. You keep pushing me for more production but you won’t give me the resources I need to do the job. I have half the number of painters I need and since you changed fibreboard suppliers from S. Lumber to Sawing Logs the quality of the raw materials has gone way down causing all kinds of rework. Or at least I think that’s what’s happening. Anyway, quality is down and nobody’s getting their vacations. And Rem the marketing VP is driving me nuts by telling everybody how everything’s great in the factory. It isn’t. I’m done.”
Barry is suddenly gripped by fear. He knows he depends on April to keep the factory humming. She’s done a great job for 30 years and she usually gets along well with both management and the people on the manufacturing line. He tries his best: “You’ve been like family to us, April. You helped mom and dad start up the business when we were only making bed knobs. You know we’re coming through difficult times but business is up and we’re starting to show a profit but we can’t afford to invest in more staff or a bigger plant quite yet. There must be a way we can get by just a little longer without working people so hard. Let’s get our data scientist on it. Maybe there’s a way. Give me another chance.”
April considers it for a moment then says, “Two weeks”.
TASK 1
A file containing production data for the Sleepy factory is on the portal: “Case 4 – Sleepy Production.xlsx”. The file contains information on production for the 249 days last year the plant was in production and includes the number of finished units produced, number of units sent back for rework, and various other facts about each day. Clean the file up in preparation for analysis.
TASK 2
Think about what might going on in the plant. Think about each of the data elements you’re provided with and how they might influence production. Run exploratory plots to get “a feel for the data”.
TASK 3
Perform a multiple regression using all of the data elements as potential predictors. Be sure to split categorical variables into a proper set of dummy (also known as an indicator) variables—do not just throw everything in as-is. You may use any tool for this purpose. Carefully interpret the results of the regression and use it to form theories about what might be going on in the plant. You won’t be able to be certain but you should find clues that you can recommend investigating.
TASK 4
Write a report suitable for Barry Sleepy that provides him with concrete recommendations to improve the situation in the factory. The report must be written in an appropriate fashion for senior management. Think carefully about what you would want to see if you were Barry Sleepy, and design the report to that purpose.
Here are some tips:
- The main body of the report must not contain any statistical jargon; use of the word “average” is okay (but not mean, for example)
- Your recommendations must be supported by solid statistical work which is presented in an appendix, where precise statistical language should be used
- Use simple, well-labeled, well-explained graphs in the body of the report but only to support specific points you want to make; do not load the report up with graphics that won’t mean anything to Barry
- Your table of contents should be:
- Introduction
- Issues at the Orillia Plant
- Findings
- Recommendations
- Appendix (Detailed Analysis)
- The report must look like a professional piece of work suitable for a management audience
NOTES
- You do not need to hand in anything separate for Tasks 1 to 3. They are there only to guide your analytical work in preparation for Task 4. Your work from Task 3 should be included in the appendix and should be well-documented as to exactly what you did and why.
- There are three shift managers at the plant: Bill, Mary, and Tom. Bill is short for William.
- Be sure to check for each variable that linear modeling is appropriate (see the requirements for use of least squares in the slides and OpenIntro and apply appropriate tests before and after the regression). Be sure to remove any redundant/collinear predictors from your model(s).
- Please submit your report on paper so we can quickly mark it and return it
