12.04.2012 Statistical Cluster Conference Call

From IMarine Wiki

Jump to: navigation, search

Contents

Agenda

Time: Thursday April 12, 2012, 16:00 - 17:00 Europe/Rome

  • Import and curation => support for formats other than csv e.g. read sdmx
  • Persisting settings of curation => Towards a Work-flow
  • Codelist manager plans => Erik
  • Use of codelists in curation => What if a codelists changes? Backwards compatibility?
  • AOB

Participants

CNR: L. Candela, F. De Faveri, L. Lelii, P. Pagano

FAO: E. Blondel, A. Ellenbroek, E. van Ingen

Discussion

Import and curation => support for formats other than csv e.g. read sdmx

How dataset reading is supported by the OpenSDMX client

  • this is still based on JAXB, this does not scale to the size of the data to be managed
  • Erik will discuss with Lucio and Federico on how to improve the client (e.g. support streams)

Persisting settings of curation => Towards a Work-flow

The Data Flow Manager discussed in AOB will support this

Codelist manager plans => Erik

A Wiki page has been created by Erik http://wiki.i-marine.eu/index.php/CodelistManager

If the Code List Manager development has to be supported by CNR, FAO and BoI than we should agree on a different strategy

  • Erik says that the approach should focus on the following topics: (i) code lists, (ii) data structures and (ii) data;
  • CNR proposal
    • the "tabular data manager" should simplify the management of any kind of tabular data in a RDBMS, including code lists
    • the "data flow engine" should support any workflow
    • because of the above two facts / ideas these tools should support code list management also
  • F. Simeoni might contribute to this activity, F. Fiorellato should be involved also
  • a discussion aiming at isolating the 'code list manager' from the rest should be organised
    • the tree-oriented store and the tabular data-oriented should co-exist

Use of codelists in curation => What if a codelists changes? Backwards compatibility?

Curated Time Series are associated (linked) with a given version of the Code Lists

  • thus if the Code List is changed without producing a new version of the Code List then the Time Series is associated with the latest one

The requirement / expected behaviour should be clarified by FAO

AOB

On tickets created in the past:

  • these have been analysed by CNR (Federico)
  • tickets will be created by CNR before the end of April
  • FAO will hire a new person to liaise between data providers and data managers for ICIS
  • FAO will not produce new tickets until the exisitng ones have been reviewed

In May we will start implementing a new component that benefits from the TimeSeries stuffs to produce

  • a Tabular Data Manager component;
  • a second component is the Data Flow Manager, this will be released before the summer

Emmanuel would like to know how to access a curated time series containing geospatial information

  • He would like to know how to identify which columns have a "geo code" inside a TS;
  • He would like to know how to acces the TS;
  • Actually, which are the code list that contains geo-related stuffs?
  • Will be discussed with GP Coro 12-04-13

Agreed actions

  1. FAO and CNR should agree on how to realize an OpenSDMX client that can read huge datasets (JAXB seems to be not appropriate)
  2. A meeting will be organised to further discuss the codelist manager, probably in May in Pisa
  3. Anton would like to be in copy with tickets related to ICIS
Personal tools