November 18, 1998 1:30pm
|Members present||Rich Forrester, Lisa Jardine, Sue Schwab, Gary Dickerson, Brad Christensen, David Heise, Karen Stockton for Sharon Pittman|
|Absent members||Emilio Garcia-Marenko, Mistee Arias, Derek Bradfield, Ann Gibson, Ron Herr, Pat Mutch, David Penner, Jack Stout, Ed Wines|
|Minutes||Minutes from last week were distributed and David Heise recapped the last meeting. Minutes were accepted.|
|Review of Data Warehousing Terms||Gary Dickerson distributed the Data
Warehouse Definitions document.
David Heise - this document is intended to be a reference document for unfamiliar terms and it will be updated as necessary. Unless we can agree on the meaning of terms, we will have problems agreeing about the meaning of the data.
Gary Dickerson - last meeting some members had questions about the term business unit. Business unit may be a term that comes into focus as we proceed, especially as we address the set of questions.
Rich Forrester - Business units may go across departmental lines in order to answer questions.
David Heise - We will have a common data warehouse that supports the separate business units needs.
|Data Warehouse Readiness Assessment||Gary Dickerson distributed the results of the Data Warehouse
Readiness Assessment the committee members completed at last meeting.
Item of note: The final score was a 71, showing that we are ready for our first increments in data warehousing. The developers of this instrument have noted that this score is typical of the scores of organizations when first tested. Scores typically go down when the assessment is completed a second time. This is due to uninformed optimism vs. informed pessimism.
On page 2, the scores to take note of are the extremely high or extremely low scores. The IT readiness score has always been low. We may need to explain what were trying to do to the non-IT people and how it will benefit them.
David Heise question on page 2, there is a weighting factor with the scores? Are the differences in the column scores significant?
Karen Stockton do they typically test top-line administrators?
Gary Dickerson this assessment was completed by the committee members.
David Heise what was your question related to?
Karen Stockton I was concerned about the data entry involved.
David Heise no data entry required. The Banner data will feed the warehouse.
Gary Dickerson the data warehouse addresses easy access to the data for end users.
Karen Stockton is this compatible with ASCII data exports?
David Heise yes, it is, and it can support a number of tools.
Gary Dickerson it depends upon the end-user tools used to access the data. It should eliminate the re-work.
Karen Stockton concern about the compatibility with SPSS and such packages.
Sue Schwab we can extract data from Banner in any format you need. Please contact me.
David Heise data mining tools can help with the statistical analysis. The complexity of the transaction data in Banner is simplified by organizing the data by business questions instead of transaction questions.
Gary Dickerson this is one of the reasons we need the business users in on the planning process.
David Heise we need people to ask the type of questions you are asking in order for this to be successful. Mark Clayton invited us to Whirlpool and he said the success of their data-warehousing project is that the users drove the process and the warehouse was created to answer the user questions.
Page 3, didnt print well, but the sheet details the measurement points related to each question.
Page 4 details the high responses, the average responses, and the low responses to each question in the assessment. For the large differentials we need to examine why the disparity exists so that we can come to common understanding and expectations for the project.
Rich Forrester What was question 53?
Gary Dickerson The IT organization worked with the business units to select the warehouse development method.
We need to look at some that were very high and see why. Were going to do more analysis to generate some items to concentrate on.
|Demonstration of End-user Data Analysis Tool||David Heise demonstrated PowerPlay from COGNOS.
Rich Forrester The business units identify the questions and what they want to analyze and ITS builds the information?
David Heise yes.
Gary Dickerson it is important that we agree on the definitions of the terms used, e.g. How do you define a student? This allows you to compare apples to apples.
Gary Dickerson can you set up tools like this to do standard reports?
David Heise yes. PowerPlay has two modes, one, which supports standard reports. You can build reports and save them to run each time the data warehouse data is refreshed.
Karen Stockton when you purchase the data warehouse, do you then decide which tools to buy?
Gary Dickerson there are two main approaches: one is the data warehouse in a box, the other is best of breed for each tool you require. At The Data Warehouse Institute conference we attended, it was suggested that you purchase the best of breed tools.
David Heise explained some of the data warehouse elements and their relationship to our needs.
Karen Stockton bottom line, how many new tools must I learn? How much does it cost me as a department?
David Heise you would pay for the client software you needed. The cost ranges from $500/seat to $900/seat, depending upon capabilities.
Rich Forrester by the time we implement a data warehouse, the technology may decrease in price.
David Heise that is almost guaranteed with Microsofts entry into the OLAP market.
Sue Schwab can we import historical data?
Gary Dickerson yes.
David Heise explained the ETL (Extract/Transform/Load) tools and how they might work. He also noted the need to bring in information from Institutional Research that isnt stored in Banner.
|Brain Storming Top Level Questions||David Heise what are some of the important pressing
questions that we need to answer at Andrews University?
Gary Dickerson we will put the list out for comments, additions by committee members and update it later.
|Next Meeting||To be advised.|