Researchers in t he EDM communit y have always relied on sophist icat ed t ools t o analyze dat a and build models. As the amount of dat a t hat can be collect ed and st ored grows, the need for t ools capable of handling “ big dat a” becomes ever more prevalent . SAS® Analyt ics U is a new init iat ive for making SAS dat a analysis and mining t ools available for free t o educational researchers and inst ruct ors. T hese t ools are designed for handling very large dat a set s and can be run in t he cloud, saving researchers valuable t ime and resources. Furt hermore, SAS Analytics U provides a communit y of SAS educat ors and learners to share resources and informat ion about SAS t ools and t echniques. T his t ut orial aims t o int roduce researchers t o t he t ools available t hrough SAS Analyt ics U and how t hey can be applied to the field of Educat ional Dat a Mining. We will provide an overview of t he SAS archit ect ure and provide inst ruct ion on t he key feat ures of each t ool in t he suit e. We will guide part icipants through examples using relevant educat ional dat a sources t o help researchers underst and how t he t ools can be applied t o their own work.
T his t ut orial will focus on int roducing SAS t o part icipants and guiding t hem t hrough t he use of t he suit e of t ools using relevant educat ional dat a set s. T he t ools t hat will be covered include: SAS® Programmi ng Language . SAS programming language is a powerful language designed specifically for int ensive data analysis. T his highly flexible and ext ensible fourt h generation programming language has a clear synt ax and hundreds of language element s and funct ions. It support s programming everything from dat a ext ract ion, format t ing and cleansing t o data analysis, building sophist icat ed models, and generat ing report s. T he SAS programming language is at t he heart of t he SAS University Edit ion t ools.
SAS® Studi o. SAS St udio is t he development environment for SAS Universit y Edit ion and runs t hrough t he web browser as well as in t he cloud. It offers a powerful GUI int erface that allows novice programmers t o int eract wit h dat a and perform analyses wit hout writ ing any SAS code t hemselves. However, the SAS code is all generat ed behind t he scenes and is visible t o help users learn.
SAS® Ente rpri se Mi ne r™. SAS Ent erprise Miner helps users st reamline t he dat a mining process t o creat e highly accurate predict ive and descript ive models based on analysis of vast amount s of dat a. It includes innovat ive algorit hms in the areas of st at ist ics and machine learning t o enhance t he st ability and accuracy of predictions, which can be verified easily by visual model assessment and validation. Users build process flow diagrams that serve as self-documenting procedures. T hese diagrams can be updated easily or applied to new problems without starting over from scratch. In addition to process flow diagrams, Enterprise Miner provides a programming interface for advanced users. Enterprise Miner allows integration with open source software for data manipulation and model comparison, the open standard PMML, and databases for scoring models without data movement.
Additional SAS tools that may be covered if it is of interest to the participants include t ools for time series analysis, forecasting, matrix manipulations, and advanced statistics.
Educational data miners rely on computational tools to understand and explore their data. T hese tools must be robust and flexible in order to allow for innovation. T hey must be able to handle ever increasing amounts of data. Ideally, they are easy to use by both programmers and non-programmers alike due to the interdisciplinary nature of this research area. Finally, most researchers rely upon tools that are freely available and do not require excessive resources.
SAS University Edition is a new option that addresses many of these needs. T his suite of powerful SAS software was made available to all learners for free in May of 2014. SAS Enterprise Miner, T ext Miner, and Forecast Server have been available through SAS OnDemand for Academics since late 2010. However, the biggest barrier to adopting new tools is learning how to use them. SAS Analytics U is a community centered around these free offerings and is designed to support SAS learners and educators. T his tutorial seeks to introduce participants to these resources and suite of tools and demonstrate how they can be applied to EDM research. T he goal is that participants will be able to add another set of tools to their every growing toolbox for conducting EDM research
T his tutorial will be divided into four sessions, each covering a specific topic: