S3Cube

S3Cube is a REST method handler for data analysis.

Functionality

S3Cube can generate contingency tables (a.k.a. pivot tables) from S3Resources.

Dependencies

S3Cube uses ​pyvttbl to generate contingency tables. A modified version of pyvttbl is integrated into the S3 framework.

Pyvttbl requires the following Python modules:

• scipy
• numpy
• matplotlib

Python 2.7 is recommended for S3Cube, however it would work with Python 2.6.

URL methods

S3Cube responds to the analyze URL method, for all resources (generic method).

The following parameters are accepted:

Parameter	Explanation	required?
rows	the name of the field to be used for the table rows	yes
cols	the name of the field to be used for the table columns	no (if no cols are specified, all instance values appear in 1 column)
fact	the name of the field to be used for the instance values	no (falls back to "name" if present, otherwise to "id")
aggregate	the aggregation function	no (default: list)

rows, cols and fact support the same options as list_fields:

• fields in the table
• virtual fields in the table
• fields/virtual fields in tables linked by foreign keys ($-notation)

S3Cube supports a number of aggregation functions. The following functions have been tested so far:

Function	Explanation
list	a comma-separated list of all instance values
count	the number of instance values
sum	the sum of all instance values
avg	the average (mean) of all instance values

Examples

• ​http://localhost:8000/eden/org/organisation/analyze?rows=country&cols=sector_id&fact=name
• ​http://localhost:8000/eden/pr/person/analyze?rows=age_group&cols=gender&fact=id&aggregate=count

$-notation for references:

• ​http://localhost:8000/eden/org/office/analyze?rows=location_id$parent&cols=type&fact=name

Further Development

The following enhancements are currently under development:

• Totals row and column
• Support for JSON, CSV and XLS exports
• Support for client-side graphs representation (via JSON)