|
|
|
Predictive Modeling Case Study
Situation
A major
university wanted to build a few new dormitories in order to accommodate
the on campus housing needs of its students. The university wanted to
know what aspects of housing students wanted the most and how to predict
the popularity of the housing before it was built.
Methodology
 |
|
 |
|
 |
|
|
| •
|
Surveyed
residents in order to determine key factors which lead them
to choose the dorm they did
|
• |
Interviewed key administration officials to determine key
variables which were considered in the past
|
|
|
| •
|
Surveyed
residents to determine quantitative measures for qualitative
variables
|
• |
Obtained data for other variables by researching records and
interviewing people at the housing office,
|
| •
|
Obtained
and analyzed university records for housing preferences
|
|
|
| •
|
Used statistical
analysis to run a regression model and assign coefficients
to each variable in order to predict dependent variable, number
of students who choose a dorm as their first option
|
• |
Optimized the R2 value (accuracy of the model)
|
|
|
|
Results
| •
|
Findings
were used to design several new dormitories and to predict the popularity
of each dormitory before it was built |
| •
|
University
has built several new dormitories which maximize student preferences
in terms of campus location, room size, dorm size, price, proximity
to cafeteria, layout, etc. |
|
|
