Company: Texas Institute for Measurement, Evaluation, and Statistics (TIMES) at the University of Houston
Company background: It is the Mission of TIMES to advance scientific discovery through the development and application of measurement, evaluation, and statistical research methods. One of our primary lines of research is the development of academic, intellectual, and cognitive abilities in school age children.
Contact: Kevin Davidson
Title: Research Associate Professor
Country: United States
Award Category: Innovative Problem Solver
Tell us about the business problem you were trying to solve?
As a research organization focused on analytics, our research faculty needed easy and reliable access to analyzable data sets. With numerous projects going on at the same time and new projects starting every year, TIMES faced disparate data sources where each research project resulted in a data silo that was clean but idiosyncratic, limiting our ability to leverage multiple projects addressing the same or related problems. The development of the warehouse has allowed us to use the data from multiple projects to ask questions that could not be answered with a single project.
Figure 1. The Problem
How did you use SAS to solve that business problem? What products did you use and how did you use them?
TIMES has a number of data scientists and programmers with strong SAS background. Together we decided to address the aforementioned obstacles by creating a data warehouse within SAS. SAS Data Integration Studio creates project specific job flows that utilize nodes of reusable code. Stored processes are easily run via the SAS Information Delivery Portal, which also provides links to SAS Visual Analytics reports.
Figure 2. Simplified Data Mart at the project level
Figure 3. SAS Data Integration Studio project cleanup job
Figure 4. SAS Information Delivery Portal showing categorized stored processes
What were the results or outcomes?
Researchers can use a custom built extraction tool (a parameterized stored process) allowing for selection not only of the project(s) of interest, but also the types of data (e.g. individual test item responses, raw test scores, standardized scores, demographics) and the data layout (e.g. one-record per subject or a repeated measures layout) conducive to the requirements of planned analyses.
Figure 5. Warehouse advantages
Why is this approach innovative?
University-based education researchers tend to view research project databases as stand-alone entities. We approached the problem differently, taking a data warehousing perspective. In this way, we were able to standardize, simplify, and make reusable components throughout much of the data cleaning process. This has paid huge dividends in freeing up data scientists to focus more on theory and methodology and far less on data preparation. Our approach is appropriate for any researcher carrying out systematic research where integration across projects opens the door for new questions and new methods. Furthermore, it complies with the Open Science standards. Education data warehouses are not just for states and school districts!
