The use of SDMX at the ECB

1
The use of SDMX at the ECB

• Xavier Sosnovsky
• European Central Bank
• Bonn, 2009-04-02

2
Overview

• Data exchange SDMX-EDI
• Data storage SDMX Information Model
• Data dissemination SDMX-ML

3
Data exchange SDMX-EDI

• 27 NCBs
• 7 NSIs
• Eurostat
• IMF
• BIS
• Using SDMX-EDI

4
Overview

• Data exchange SDMX-EDI
• Data storage SDMX Information Model
• Data dissemination SDMX-ML

5
Data storage SDMX Information model

• Data stored in FAME and Oracle databases
• Based on the SDMX information model (concepts,
  code lists, key families, dataflows, etc)

6
Overview

• Data exchange SDMX-EDI
• Data storage SDMX Information Model
• Data dissemination SDMX-ML

7
SDMX-ML on the ECB websites

• Export functionality on the ECB Statistical Data
  Warehouse
• Bilateral exchange rates euro area yield curves
• Publication of national contributions
• ECB visualisation tools

8
The ECB Statistical Data Warehouse

• Full support of extractions in SDMX-ML format

9
Bilateral exchange rates and euro area yield
curves
Extractions into SDMX-ML format
10
Publication of national contributions (1)

• The problem Need to improve the accessibility of
  ESCB statistics (user request).
• Time-consuming task
• Accessibility issues
• The solution To present euro area aggregates and
  national contributions in one table,
  simultaneously on the ECB and participating NCBs
  websites.
• The benefits
• Increased user-friendliness
• Consistent data and presentation
• Minimised maintenance work

11
Publication of national contributions (2)

• Technical overview
• SDMX standards SDMX-ML 1.0 Compact format data
  files published on the ECB website
• XSLT a programming/scripting language (i.e.
  Java, PHP, etc)
• Multi-channel output (HTML, CSV, etc)

12
Publication of national contributions (3)

• NCBs implementations Own added values.
• Localization
• Database with wizard and personalization features
• Display of metadata
• Additional download formats

13
Publication of national contributions (4)

• Benefits
• The ESCB statistics sites appear as a
  co-ordinated network with an entry point in all
  NCBs
• Users do not have to leave the NCB website to
  consult ESCB data
• Look and feel of NCB website is preserved
• Translation into national languages

14
ECB visualisation tools (1) Current applications

• ECB dynamic visualisation tools use SDMX-ML as
  format for data feeds. Examples
• The bilateral exchange rates graphs
• The euro area yield curves
• The inflation dashboard

15
ECB visualisation tools (2) An SDMX
visualisation framework

• Develop libraries that can be used to build
  visualisation tools for statistical data and
  metadata
• Written in ActionScript 3 (Flex/Flash
  applications)
• Based on the SDMX information model v. 2

15
16
ECB visualisation tools (3) An SDMX
visualisation framework
ECB SDMX Framework
SDMX-ML Readers
Read
Translate
SDMX Information model
Use
View
Access
Users
Visual display
16
17
ECB visualisation tools (4) An SDMX
visualisation framework

• Open source project
• ECB, NY FED, BOC, OECD
• Code hosted and publicly available on Google Code
• Strict commit policy and change management
  procedures
• BSD license
• Roles
• Leadership roles (Project Manager, Project
  Advisor and Project Architect)
• Committers three individuals
• Contributors contributions are welcome from all
• Users the code is available to all

17
18
ECB visualisation tools (5) Planned work

• Generalise the use of this technology to other
  sections of the ECB website (e.g. effective
  exchange rates, HCIs, Eurosystem joint
  dissemination sections, etc.)
• Replace charts of the ECB Statistical Data
  Warehouse with similar interactive visualisation
  tools
• Add new statistical domains to the ECB dashboard

19
SDMX Web services on the SDW (1)

• Available 2009 Q3
• Possibility to get all statistical data stored in
  the SDW
• Support for the SDMX-ML Query format
• Possibility to get all the data or updates and
  revisions only
• Possibility to retrieve data snapshots
• Possibility to filter data by dataflows,
  datasets, dimensions, attributes, date ranges,
  etc
• Data returned in SDMX-ML Compact Data and SDMX-ML
  Generic Data formats (v. 2)

20
SDMX Web services on the SDW (2)

• Possibility to retrieve data structure
  definitions, code lists and concept schemes
• Possibility to retrieve the category schemes,
  organisation schemes and dataflow definitions
  (including constraints)
• Metadata returned in SDMX-ML Structure format

21
SDMX Web services on the SDW (3)

• Possibility to submit subscription requests and
  receive update notifications, when specific
  actions occur on data and metadata
• Possibility to receive notifications via email
  and via an HTTP request