11.07.2013 ICIS and validation tools in iMarine
From D4Science Wiki
Conference details
- Date: 11/07/2013
- Where: Google Hangout
- Participants:
- Yann Laurent (FAO)
- Anton Ellenbroek (FAO)
- Pasquale Pagano (CNR)
- Leonardo Candela (CNR)
- Luigi Fortunati (CNR)
- Topics: iMarine, VALID initiative, validation, statistical data
Agenda
- ICIS new version presentation for CNR and discussion with FAO
- DG-Mare VALID web service needs presentation by FAO and discussion with CNR
- StatiCube bundle components discussion:
- Collect: presentation of DCF opportunity – presentation of data collection tool by CNR
- Process: ICIS – other needs ?
- Analyse: ICIS – other needs ?
- Disseminate: any planed tool ? StatBase (or StatBase like) as candidate for implementation in iMarine
- Archive: iMarine infrastructure
(the above phases are from UNECE Generic Statistical Business Process Model. We could also use as reference DDI v3 that might be more explicit as it talks about data collection instead of collect)
Notes
- ICIS new version presentation for CNR and discussion with FAO
- LF presented activities of CNR (slides from last TCom) on timeseries management: system architecture, features, status.
- LF: When a user upload a csv file a generic tabular data table is generated. From that point the user can transform the table into a dataset or a codelist or another type of table by performing transformations. Import from SDMX should directly create dataset or codelist tabular data tables.
- YL: If 2 columns both refer to a CL, these must be managed, and also subsets of CL’s in datasets must be managed, especially when disseminating data. YL will work on examples and produce tickets later.
- YL: DG MARE wants to track corrections. Add metadata to specific measures / tuples of a dataset.
- LF: We cannot yet add a metadata to single table cell (e.g. specific observation measure) or individual tuples. We'll consider this use case.
- LF: Developing SDMX data client, JSON import/export, and transformations
- YL: Will provide more info with ticket #465 (Column merging using algorithm).
- LF: At first we can provide simple means to merge two columns. After that we can work on providing a language to be used in order to achieve greater flexibility for this task.
- YL: Duplicate line removal starts often when deleting a column. Some transformation should trigger other transformations.
- PP: Both Union and Aggregation will be supported, as well as import and update rules.
- LF: A gazetteer will be added. 1 column will be linked to 1 CL, if a mapping is not possible, a new CL can be generated that can be managed a private list. If data are in a hierarchical CL, then a solution may be found.
- LF: Data Analysis: JPivot can be used with Mondrian and MDX to analyse data. Other tools like saiku have the same capabilities and can be integrated with the GWT UI.
- LF: First mock-up UI release is planned by July for feed-back.
- PP: Most of basic the back-end components are ready. Import data workers, Cube Manager, query module, service framework.
- DG-Mare VALID web service needs presentation by FAO and discussion with CNR.
- Asks for manager level validation tools where
Actions
- Have another call next week before VALID meeting.
- Yann will send to CNR the document presented about VALID initiative.
- CNR will read the document and provide feedback at the next call conference (planned for next Thursday).