Data Warehouse and Reporting Panel Discussion
Panel Participants University of Northern Colorado Bret Naber – Director of Applications and Database Development Colorado School of Mines Tom Boyd - Dean of Graduate Studies Colorado Community College System Julie Ouska - Information Officer / Vice President
UNC Reporting Architecture ODS 3.1.1- Oracle 10.2 - 42 gig EDW 1.4 Nightly Refresh average – 1-2 hours Full Loads – 20-25 hours Banner 7.3.x - Oracle 10.2 - 60 gig Webfocus 7.6.6 Developer Studio Report Caster Resource Analyzer AD Authentication
More information All Faculty/Staff have accounts Three Tiers of users Request access to “Views” Three Tiers of users Report viewer Report writer Report developer
UNC - Project Goals Create a central repository of all critical institutional data that is available to all academic and administrative units Validated data Appropriate data security
UNC - Project Goals continued Provide relevant , consistent and timely reporting and analysis to decision-makers at all levels of the university Single source of the truth Standardized data definitions Relevant data collected Decentralized skilled report writers Identification of key performance indicators Centralized analysts of complex information
UNC - Project Goals continued Consolidate and streamline existing reporting processes Effectively use specialized skills of complex analysts Leverage business process and subject matter expertise of report writers in administrative areas
TemporaryProject Manager Project Leadership TemporaryProject Manager Decision Support Project and Information Management Team IT ADDS Team System Adminstrator OBIA Strategic Analysts College Power Users & Report Writers Other Academic Power Users & Report Writers Business System Trainer/ Report Director Temp Report Coordinator - Peggy Hoff Administrative Data Stewards (incomplete list) Graduate Robbyn Wacker Report Coordinator Power Users & Report Writers Auxillary Tobias Guzman Financial Aid Donni Clark Student Nolan Oltjenbruns Admissions Chris Dowen HR Marshall Parks Finance Paul Squillace Various Cross-Functional Project Teams
CHECO Building Your Own Data Warehouse September 23, 2008
I’ll go on the record that I don’t recommend the “Build It Yourself” approach. But if you want to go that direction, I have some suggestions. Fall CHECO Conference 11/11/2018
Start by thinking about the end result first Institutional research? Operational reporting? Longitudinal analysis? Student / Financial Aid? Finance? HR? Fall CHECO Conference 11/11/2018
Random Thoughts The more data required, the more complex the infrastructure and technology becomes Analysis over time requires the ability to freeze data snapshots It is not always obvious where the data resides in Banner – you will guess wrong a lot. Think about ETL upfront. Lots of data will probably require an incremental refresh later. Very difficult to retrofit. Not an IT project – executive leadership is required. Functional user input also required. Banner must be the authoritative data source – don’t “fix” stuff in the DW that is wrong in Banner, fix the data/process in Banner. Start small and expand on success Fall CHECO Conference 11/11/2018
Infrastructure is critical If the extract, transform & load process does not work well, your staff will spend all their time working on ETL, and not on developing views and reports. Make it easy to add new data Put data in the DW once, use it many times Agree on the transformational values such as FTE, full-time, part-time, etc., up front and across the institution Fall CHECO Conference 11/11/2018
Success factors Discrete project with a reasonably sized data set Go for the 80/20 rule Look at the reports that are used over & over – variation on a theme Use a web-based, click & point reporting tool for most users Expand use slowly, use focus groups to improve data and delivery Validate data very carefully & thoroughly Fall CHECO Conference 11/11/2018
Additional thoughts or questions?