Transcript Otsikkodia - Statistica

Multichannel publishing of statistics
(electronic publications and database) Finnish experience
Seminar on dissemination of statistics and
launching a new web solution in NBS, Moldova
Chisinau, 27. May 2010
Markku Huttunen
[email protected]
Changing over from a printed statistical system to
an electronic one
Statistical information has been reported, released and
disseminated in the form of printed publications
 Statistics Finland’s oldest statistical publication series has
been published for over 250 years (>)
 Internet has put the traditional way of reporting, publishing
and disseminating statistical information in turmoil
 Traditional form of publishing is disappearing
(=tables and the information needed in their interpretation in
one volume in a publication series of official statistics)

Markku Huttunen
27.5.2010
2
Common Structure of Statistical Information
(CoSSI)
The point of departure in the CoSSI was an (infological) analysis of the
information being considered.
 The conclusion from the analysis was that although in practice the
definition of statistical information has varied according to a given
situation and application, in reality statistical information has a certain
simplifiable and acceptable universal structure.
 The CoSSI describes the general structure that is not dependent on the
situation of the statistical information presented in differing formats.

=> CoSSI defines the structures of statistical data, metadata and
publications (>)
Markku Huttunen
27.5.2010
3
Dissemination process - Office97

Figure: Harri Lehtinen
Markku Huttunen
27.5.2010
4
XML based dissemination process - XML
and PC-Axis

Figure: Harri Lehtinen
Markku Huttunen
27.5.2010
5
Archiving of an electronic statistical publication
How will the statistical information published today
only in electronic format be available in 10 or 30
years’ time?
 In CoSSI model the XML file of the publication contains the
information content, the statistical metadata and the
publication metadata in the desired languages in one
XML file -> archived in the eXist XML database
 Keeping published publications permanently accessible to
users is an essential element of a good information service
in time (>)

Markku Huttunen
27.5.2010
6
Roles of an electronic statistical publication and a
table database (1)
Tables in databases are different than tables in
publications (printed or electronic)
 Database tables are not archived as tables published in
printed, HTML or PDF publications are
 In database tables, data on older observations are always
published according to the latest situation
 Linking an electronic publication to the related database
tables is a problem: data only correspond with each
other at the moment of publishing

Markku Huttunen
27.5.2010
7
Roles of an electronic statistical publication and
table database (2)
At Statistics Finland, tables in PX-Web database are not linked
to any other background database: PX-Web database builds
up on the server from published PC-Axis files
 In the XML publishing system PC-Axis files published in the
PX-Web server are saved into a directory of statistics


Publications and related database tables are archived

The electronic publications and the related database tables (in
PC-Axis format) are archived unchanged -> two advantages:

1) it makes sense to link the tables published in the database to the publication

2) an archive is created for the tables published in the PX-Web database (>)
Markku Huttunen
27.5.2010
8
New practices for releasing statistics
Statistical releases 650 / year (FIN, SWE), 350 (ENG)
 The objectives and requirements of the XML publishing
process under construction were taken into account



e.g. archiving and multichannel distribution
Statistics Finland’s XML-based publishing system


over one thousand XML-publication titles are produced 2007-2010
in May 2010 from 200 statistics 150 use the XML-publication process
Markku Huttunen
27.5.2010
9
Redesigning of publications

Old publishing process (Office97)



one statistics publishes 1) a statistical release (minimum) 2 ) printed
publication 3) tables in statistical database (StatFin/StatLine)
all produced using different production processess, different
dissemination channels and with different content
New publishing process (XML and PC-Axis)


one statistics publishes 1) a statistical release (minimum) 2) electronic
publication (+ print) 3) tables in statistical database (StatFin/PX-Web)
statistical release and publication (1 and 2) have merged into one entity
Markku Huttunen
27.5.2010
10
Basic publication from statistics
The role is to release basic data on the official statistics
 An entity that describes a single set of statistics at a certain
reference time



statistical release, a more extensive text section, table and figure
annexes and a standardised quality description
The objective is to transfer the basic publications from
statistics comprehensively into the new XML-based
publishing process
Markku Huttunen
27.5.2010
11
Redesigning tables
The role of tables in a basic publication from a set of statistic is
to describe the key points in a compact way
 Even though PDF publications are also created for multichannel
distribution, the main principle is “electronic first”
 Tables must function properly in HTML format, so they have to
be small and deal with only one subject


usability and space (>)
Markku Huttunen
27.5.2010
12
Browser view
Vertical A4
Printed Statistical Yearbook

Figure: Jaakko Laakso
Markku Huttunen
27.5.2010
13
Switching over from old-style printed publications to
electronic publishing
Majority of the tables in publications, which are larger
than a single A4-sized page, will be transferred to the
database
 The basic publication from a set of statistics is divided into
electronic content published online, with two different
formats and user interfaces (HTML and PDF) + tables
published in a statistics database

Markku Huttunen
27.5.2010
14
Sources

Common Structure of Statistical Information (CoSSI)
<http://www.stat.fi/cossi>.
Markku Huttunen
27.5.2010
15