toaster version 0.4.1 - first R package for Aster - is available from CRAN

The best minds from Teradata, our partners, and customers blog about relevant topics and features.
Teradata Employee

With the release 0.4.1 toaster - first R package designed and developed for Aster - received its latest update that includes comprehensive support for kmeans clustering, new utility function to compute NULLs, and support for the latest version of ggplot2 (2.0.0). It's available from CRAN or by running this in R:


If you didn't follow the package's timeline before 0.4.1 then here is a high-level review of toaster features:

  • summary table statistics with function getTableSummary;
  • quick access to standard statistics and their visualizations with showData;
  • family of compute functions to perform computations, processing and aggregations on "big data" in-database and return results to R;
  • family of create functions to quickly turn results of compute's into effective visualizations;
  • the following in-database functionality is available via compute functions:
    • arbitrary aggregates,
    • aggregates for barcharts,
    • aggregates for heatmaps,
    • correlations,
    • histograms,
    • linear regression,
    • kmeans clustering,
    • silhouette clustering technique,
    • percentiles,
    • sampling,
    • TF and TF-IDF (text analytics),
    • Ngram and token text parsers;
  • the following visualizations are available via create functions:
    • boxplots,
    • bubblecharts,
    • heatmaps,
    • histograms,
    • population pyramids,
    • slopegraph,
    • geographical maps,
    • wordclouds,
    • specialized clustering plots;
  • support for maps includes geocoding and automatic map location;
  • utility functions to work with Aster tables and data types.

toaster is open-source public project on github:

Its wiki page makes sense as a place to get started with toaster: Home · teradata-aster-field/toaster Wiki · GitHub

You can also refer to this and this legacy pages for more practical examples.

Special thanks to Mike Riordan, Chip Starkey, Steve Noynaert and Nick Switanek for their help with package development, testing, and valuable feedback.

You can submit bug reports, enhancement and feature requests on the project site: