Cape - a system for explaining outliers in aggregation results through counterbalancing.
Cape () is a system for explaining outliers (surprisingly low or high) aggregation function results for group-by queries in SQL. The user provides the system with a query and a surprising outcome for this query. Cape uses patterns discovered the describe trends that hold in the data in aggregation to explain such outliers.
Cape requires python 3. Install with pip:
pip install capexplain
Install from github
git clone firstname.lastname@example.org:IITDBGroup/cape.git capexplain
python3 setup.py install
Cape provides a single binary
capexplain that support multiple subcommands. The general form is:
capexplain COMMAND [OPTIONS]
Options are specific to each subcommand. Use
capexplain help to see a list of supported commands and
capexplain help COMMAND get more detailed help for a subcommand.
Cape currently only supports PostgreSQL as a backend database. To use Cape to explain an aggregation outlier, you first have to let cape find patterns for the table over which are you are aggregating. This an offline step that only has to be executed once for each table (unless you want to rerun pattern mining with different parameter settings).
capexplain mine [OPTIONS] to mine patterns. Cape will store the discovered patterns in the database.
To explain an aggregation outlier use
capexplain explain [OPTIONS].
Cape is developed by researchers at Illinois Institute of Technology and Duke University. For more information and publications see the Cape project page http://www.cs.iit.edu/~dbgroup/projects/cape.html.
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.