2. December 2015: Interagency and Expert
Group for SDG Indicators (IAEG-SDGs)
submits its report to the 47th United Nations
Statistical Commission
March 2016: UNSC agrees the SDGs indicator
framework
April 2016: UNSD requests SDGs data from
specialized agencies
May 2016: SDGs indicator data begin to arrive
18 July 2016: SDGs report and database due
to be released
3. May 2016: SDGs data
begin to arrive
18 July 2016:
Release database
?
4. April 2016: No certainty over indicators and
disaggregation to be received even in the short
term; relative clarity only reached in June
To be completed in less than two months:
◦ Receive and catalogue the data
◦ Carry out data entry
◦ Validate and release the data
To be developed:
◦ SDGs database
◦ Data entry tools
◦ Data validation tools
◦ Data dissemination
5. Because of time pressure, traditional data
systems development would not work
SDMX tools had the promise of making it
possible to configure a database system from
building blocks, with little or no software
development involved
6. MDG DSD was used as a foundation for the
SDGs database
Concepts and dimensionality were retained
Code lists were replaced to match the SDGs
dataset
The DSD had to be updated multiple times
throughout the exercise as data arrived and
codes had to be added for indicators and
their breakdowns, units of measure, etc.
7. IStat Loader was used to create the SDGs
database
◦ Automatically creates a database from a DSD
◦ Automatically configures mappings in SDMX
Reference Infrastructure for dissemination
The internal SDGs DSD was fed to Istat Loader
to create and update the database
8. Eurostat’s SDMX Converter used for data
entry
Excel data entry spreadsheets were created
and mappings configured between their cells
and the SDGs DSD
9. The data entry team copied SDGs data from
loosely formatted incoming Excel files to the
data entry spreadsheets and added codes for
series, dimensions, etc
SDMX Converter was used to retrieve data
from the Excel data entry spreadsheets and
format it as SDMX messages
Istat Loader was used to upload the SDMX
messages to the database
10. Eurostat’s SDMX Reference Infrastructure
(SDMX-RI) was used to retrieve data from the
database
Istat loader configures SDMX-RI mappings
automatically
11. SDMX-ML Data retrieved from the database was
then converted to JSON, from which presentation
was built for the SDGs Data and Visualization
Platform
Dissemination at UNSD’s UnData Platform was
configured by adapting procedures developed for
data exchange with UNESCO, which used a DSD
with the same concepts as MDG
SDMX dissemination through UnData API will be
available shortly
◦ And will then be used to power the Data and
Visualization Platform
13. Issues with special characters and extended
Latin characters in the tools used
◦ Sometimes took a lot of effort to investigate and
resolve
Updating the database from an updated DSD
was tedious and error-prone
These are teething problems, likely to be
overcome soon
14. Data processing and content validation had to
be done in Excel, outside of the system
◦ Validation and Transformation Language will
address this issue
Conversion from SDMX-ML to JSON had to be
developed manually using a non-standard
JSON format
◦ Will be resolved when SDMX-JSON support is added
to SDMX-RI
15. A high-profile database system was created
in a few weeks and populated with over
300,000 observations
Technology already exists to set up data
systems from building blocks
Dramatically lower cost of development,
dramatically higher productivity
Promise of more to come