CIS 624 Spring 2008 Assignment 5 – Cognos
OLAP Software
Due: Start of Class on 05/01/08
– at the start of class – but try to do in class on 4/24
Task:
There is a multidimensional data cube
available called “OceanCity” on a La Salle computer science www server. Access
it and the Cognos PowerPlay software by accessing http://www.lasalle.edu/~redmond/teach/624
then looking for “research materials”, then for Cognos software. Following that
link will bring you into PowerPlay with a choice of data cubes. Use the “OceanCity”
link for this assignment!! (Note that in Cognos PowerPlay, a relational
dimension (e.g.) is broken down if there is more than one hierarchy within it
(e.g. Bedrooms, Bathrooms, etc)
There are 7 tasks described below.
Hand in a printout for each (or an electronic copy of a screen shot). Many also
ask questions of you – answer those. We will have approximately ˝ of the
evening for you to work on the assignment – please ask questions if it is
unclear what a task is (generally they are looking into aspects of the beach
rental business that seemed interesting – data is entirely made up however!!!).
Generally, if the description of a table says “show X by Y by Z”, I want X as
the measure, Y as the row headers, and Z as the column headers. Please put your
name on all hardcopy printouts.
Details:
- Hager is the biggest renter of Ocean City beach houses. Their
“southend” office is their biggest renting office. Drill down to agents
within than office are down. Show number of rentals by renter state by renting
agent (within Hager Southend office). Question: What three
agents had the most rentals (in order)? (sorting by total number of
rentals could be valuable)
- Returning to the Office level (still with Hager), we’re
investigating rentals by number of blocks from the beach. Show a clustered
bar graph illustrating Hager’s rentals by office by blocks from the beach.
Make sure that the graph makes it easy to compare rentals by the number of
blocks from the beach for a given office. Question: For
each Hager office, how many blocks from the beach was the least popular?
- Starting from Scratch with a new line of inquiry,
We’re investigating new houses right at the beach – we want to filter by
new condition and blocks to the beach of zero. Show the number of rentals
by number of bedrooms by type of unit (duplex, single, townhouse, condo) .
Sort by total rentals across all unit types. Question: What
are the two highest selling (renting) number of bedrooms? (calculate this
by hand based on info in the table).
- Following up on the previous query, show rentals as
a percent of column total (instead of as values) Question: Among
the types of units, which one type has the highest percent of
its rentals being four bedroom units? (answer should be one type of
unit!)
- Starting from Scratch with a new line of inquiry, Drill
down on Berger offices. Filter to houses with Central Air Conditioning. Show
the number of rentals by Renter County by Berger Office locations. Do
80/20 suppression to focus on only counties with larger numbers of
rentals. Question: Which three counties produced the most
rentals at the southend office (in order)? (sorting by southend rentals
could be valuable)
- Starting from Scratch with a new line of inquiry, Filter
to rentals of houses with 4 bedrooms, 2 baths, 1 block from the beach,
with central air conditioning, non-smoking, in very good condition. Do
80/20 suppression. Show a line graph, showing number of rentals for each renter
state (one line for each) for each of the 5 years in the history. Make
sure your graph emphasizes trends over time. Question: Which
year seemed to have more than usual rentals from Pennsylvanians and less
than usual rentals from New Jerseans? (among houses meeting these
criteria).
- Starting from Scratch with a new line of inquiry, Show
rentals by Ocean (front, View, non) and condition (nested!) by
Year. Show an additional column with the changes from 2001 to 2005 (an
increase should be positive). Question: What ocean/condition
combination had the largest increase in rentals between 2001 and 2005?