Data SGP
Data sgp is a package that provides classes, functions and data for conducting student growth percentiles and percentile growth projections/trajectories using large scale, longitudinal education assessment data. It utilizes quantile regression to estimate the conditional density associated with each student’s achievement history and uses the resulting coefficient matrices to calculate percentile growth projections/trajectories needed to reach future achievement targets.
The current SGP is calculated for students who have taken at least two tests in different testing windows (testing windows do not need to correspond to a school or district’s school year). The calculations include compensation for the error associated with estimating a student’s prior performance by comparing their current test score with an estimate of their previous test score. The lower level function studentGrowthPercentiles performs this comparison. The SGP package includes higher level functions, called wrapper functions, that simplify operational SGP analyses source code.
The SGP package also provides a dataset that contains state specific meta-data in addition to the data used for student growth percentiles and projections/trajectories. This dataset is useful for researchers looking to replicate SGP analyses and a variety of other educational research. It is available via a web interface and can be downloaded for offline use.
Although SGP researchers are working with an unprecedented amount of data, in terms of the size of other datasets that may be considered ‘big data’ this is still relatively small in comparison. In fact, SGP data is a drop in the ocean compared to, for example, a study of Facebook interactions.
While the SGP package is designed to work with big data, it does not require a complex schema or a custom database management system. In fact, the majority of data sgp is in long format and is manageable by existing tools such as rdf, pgsql or pdb.
Aside from storing the raw data, the SGP package is also designed to provide tools for managing it and creating visualizations. Among these are a spreadsheet that allows for easy sorting, filtering and visualization of the data as well as a GUI for generating reports, plots and other graphics from the raw data. The GUI is designed to be easily extensible and can be customized for a researcher’s needs. It is currently available in a beta version for download at the SGP website. A new release is planned for the near future that will include a GUI for importing and exporting data as well as support for additional file formats. This will make the SGP package a much more usable tool for researchers of all kinds. This will be especially important for those who need to share their results with other researchers and policy makers. The GUI will also be useful for developers who want to incorporate the SGP package into their own applications. SGP will continue to be a free, publicly accessible software package. However, it is important that its users contribute to its development by providing feedback on new features and functionality as well as contributing bug reports and fixes for any issues encountered.