Vedrana Milosevic, Swedish Social Insurance Agency, vedrana.milosevic@forsakringskassan.se

Swedish Social Insurance Agency is one of 29 agencies in Sweden that are responsible for official statistics. In 2017 we started an agile project with the aims to make our official statistics more accessible and clearer. In this project we have explored unconventional statistical presentation such as a statbot (statistical robot similar to chat functions on various webpages) and storytelling as well as more traditional ways of presenting statistics, through key figures and statistical databases.

In February 2021 we released an improved webpage www.fk.se/statistik which relays on common statistical presentation that our users are familiar with, such as key figures, statistical database and descriptive statistics. It is not the result that is new in any way rather the way in which we came to the result. Through agile working methods we have worked closely with user experience (UX) designers, copywriters, communication specialist and IT to develop and user test different ways in which statistics is presented.

By interviewing current users, two user groups were identified, the explorers and the experts. The characteristics of the explorers are that they aren’t familiar to social insurance and/or statistics. The experts on the other hand are accustomed to social insurance and/or statistics.

One of many things the user tests showed was that few knew that the Swedish Social Insurance Agency is a statistics producer. As a consequence, users google to find the statistics they are looking for. One way in which we have worked to get better google results is through descriptive statistics. These descriptive statistics are presented in a concept we call Commented statistics, where we address the explorer, explaining different concepts of social insurance as well as answering some of the most common statistical questions.

Continues user testing is needed to keep our statistics accessible and clear to the explorers and experts.

Keywords: user testing, user groups, agile project management, statistical presentation

Session 3.1. Work for accessible statistics

When: Tuesday, August 23 at 13:00 - 13:55

Download presentation

Ragnhild Rein Bore, Head of user insights and web development, Statistics Norway, ragnhild.bore@ssb.no

Trine Ånerud, senior adviser, User insights and web development, Statistics Norway, trine.anerud@ssb.no

For many organisations with large, complex and information heavy websites, the EU Web Accessibility Directive and the Web Content Accessibility Guidelines (WCAG) can strike fear into the hearts of the most accomplished webmasters and web editors, project managers, heads of communications etc. Or perhaps more commonly, the urge to bury your head in the sand, hoping no one will look in your direction and praying no one external will audit your services.

In Statistics Norway, we had a few idealistic employees in the communications department who tried to the best of their ability to make sure we followed the WCAG guidelines on our website and in our statbank. But there was little real understanding of the importance of accessibility among directors and head of divisions, myself included. We paid lip service to the need to follow the guidelines, but it was done with little real follow up in terms of:

• Increasing our competence in the field of accessibility and knowing when something is good enough.

• Building robust systems, processes and guidelines

We would argue that this situation is pretty common in many national statistical institutions. We know the directive is there, our user experience people know the WCAG guidelines, but we are slow to understand that it means we have to change as an organisation. Accessibility needs to permeate the entire organisation, and it needs to be mandated from above.

In Statistics Norway we are in the middle of building such a necessary culture of accessibility across the organisation. Everyone – from the data collectors, statisticians, developers, UX designers to the lawyers, project managers and heads of divisions need to know why accessibility is so important.

In this paper we present the key initiatives taken, and I would argue that the two main reasons behind this change is:

• Empathy. The need to embrace the moral obligation of making sure everyone, including people with disabilities, are able to perceive, understand, navigate and interact with the Internet.

• The need for some professional outside to help identify where you are at as an organisation in terms of accessibility

Keywords: accessibility, user needs, organisational change, WAD/WCAG

Session 3.1. Work for accessible statistics

When: Tuesday, August 23 at 13:00 - 13:55

Download presentation

Svetlana Badina, Statistics Norway, Svetlana.Badina@ssb.no

Modern society is characterized by the wide usage of information technologies and social media among young generations. Reliable statistics are an important tool for understanding the world around us and as a basis for critical thinking. Statistical illiteracy and a non-critical approach to data, numbers and statistics can potentially affect young generations’ approaches to the future solution of global problems.

To meet this challenge, Statistics Norway has made a strategic choice to have secondary education students, as a target group for adapted dissemination of knowledge about the value of using official statistics for the society. We reach students by approaching schools and teachers.

Statistics Norway wants to encourage students to be critical to the huge volume of information they have access to, and to get familiar with official statistical sources, in an adapted manner. At the same time as tuition needs to be suited for students who have an average engagement in statistics, and for students who might find this subject difficult, school curriculum about statistical methods may be too scarce for the most engaged and curious students. Professionals from statistical agencies can assist in filling competence gaps in a rapidly changing world, and to build an important bridge between theory and practice.

Building on experience from and following the examples of assignments in the European Statistics Competition in 2021 and 2022 we have developed and started testing teaching material in survey methodology. The material includes creative assignments based on surveys being conducted in Statistics Norway. Further, Statistics Norway has increased efforts to produce teaching material about statistics and analyses for various topics, and to communicate with teachers about them.

In this article, we provide an overview of the recent achievements and current challenges in Statistics Norway’s efforts to help young generations be statistically literate.

Keywords: statistical literacy, critical thinking, survey methodology, curriculum, teachers, school students, user contact, adapted teaching material.

Session 3.1. Work for accessible statistics

When: Tuesday, August 23 at 13:00 - 13:55

Download presentation

Linn Seime Thistel, Statistics Norway, linn.thistel@ssb.no

Frode Larsen, Statistics Norway, frode.larsen@ssb.no

This year, a new Language Usage Act was introduced in Norway. For the first time, the law clearly states that the public sector must write in a plain and correct manner, suitable for the target users. For Statistics Norway, the law formalizes and strengthens what has been a priority for many years. An important part of our mission is to write in order for the public to understand and use the figures we gather and interpret.

When we write about a complicated subject, our choice of words can matter a great deal to the reader, with respect to comprehensibility of the content, and in order for the user to discover the text in the first place. The best words are not necessarily the most accurate ones, rather the words that best match the users’ natural discourse might prove expedient.

In this paper, we will look at three different categories where we aim at using the best words for the public, but regretfully without succeeding. We will look at the reasons for why our efforts were insufficient, show examples, try to determine why we did not get the results we aimed for, and for each category point to a better solution for the future.

The three categories we will examine are: Search terms, simplifying categories and overcomplicated words.

We will show how we work systematically to analyze the users, and how we sometimes discover that they have evolved ahead of us linguistically.

We will show that efforts to help the user by simplifying words and categories sometimes can turn out wrong.

We will present the challenge of lingo or complicated language that can create a gap between us and the users.

Keywords: Dissemination, words, users

Session 3.2. Communication

When: Tuesday, August 23 at 14:00 - 14:55

Download presentation

Kristin Fredriksen, Department for Communication, Statistics Norway, kristin.fredriksen@ssb.no

Nina Tuv, Department for Communication, Statistics Norway, Nina.Tuv@ssb.no

When Norway went into lockdown in March 2020, there was an increased demand from the authorities and the media for figures that could explain the economic effects. At the same time, there were several private analysts and forecasters who took part in the public debate. These relied heavily on Statistics Norway’s figures but did not present the complete overview and insight that Statistics Norway could. The Department of Communications saw the potential to be an important contributor to the public debate. At the same time, the Department for National Accounts struggled with how they should communicate the huge uncertainty in their produced figures, as a lot of the data base had vanished overnight. There was an obvious need for cooperation and benefitting from each other’s expertise.

The cooperation between the two departments had a rough-ish start. The importance of the matter, however, finally pushed us to find a way to work together. After some time, the value of this cooperation became crystal clear to both departments. We managed to communicate our figures not only to our expert users, but also to the general public who now could understand the economic effects of the pandemic. A great deal of effort was put into storytelling towards target audiences and media appearances.

In many ways, this was the start of a new working method for the communication advisors. In this paper, we will present our experiences on how to succeed in the cooperation between statisticians and communication advisors – two professions that traditionally have very different views on dissemination to begin with – and how both can end up content with the result. And, of course, how this produce better results in reaching our different target audiences. The paper presents the Department of Communications point of view.

Session 3.2. Communication

When: Tuesday, August 23 at 14:00 - 14:55

Download presentation

Peter Tibert Stoltze, Head of Methods and Analyses, Statistics Denmark, psl@dst.dk

In 2015, Statistics Denmark (DST) started compiling and publishing a series of crosscutting analyses under the headline “DST Analyses”, and since the beginning more than 100 analyses have been published. (As of January 2022, the count is 137.)

A DST analysis is a relative short (typically 5-10 pages) description of a phenomenon in society of general interest that typically is also covered by the official statistics program. However, in an analysis we explore the phenomenon under study from multiple points of view and supplement the usual statistical tables with more elaborate and explicit interpretations. Thereby we deviate in a substantial way from the traditional disseminations, which Statistics Denmark have published since its inception in 1850.

The talk covers various aspects of DST Analyses including examples of a few specific analyses that for one reason or the other have had special impact. We will also describe the process involved in deciding on topics for future analyses, the different collaboration models used for producing the analyses (including collaboration with people not employed at Statistics Denmark), and the review process undergone before publication. The publication process can vary to a certain degree, as sometimes a press release follows the analysis and some kind of infographic is produces for social media (primarily Twitter and LinkedIn). We follow closely the number of downloads and general reach by press surveillance. This has revealed a surprisingly long period of interest for the analyses.

Although the decision to start producing analyses at Statistics Denmark was a bit controversial, it has proved to be a success and is now an accepted part of Statistic Denmark’s activities. A very important result of the analyses is that the staff now has a well-established medium the share their deep insights in various domains that goes beyond tables and graphs.

Keywords: Analysis, dissimination, crosscutting

Session 3.2. Communication

When: Tuesday, August 23 at 14:00 - 14:55

Download presentation

Mette-Line Jakobsen, Statistics Denmark, mjk@dst.dk

Statistics Denmark uses a three-level topic taxonomy in our dissemination of statistical products – e.g. in our database Statbank Denmark, at our website dst.dk, in newsletters and on other communication platforms. The taxonomy is a key tool for our users to find relevant content at Statistics Denmark.

However, inquiries from users and a number of user surveys have shown that a large number of users had difficulties navigating the taxonomy. To address this issue, the Communication division in Statistics Denmark recently carried out a project developing a new and more intuitive taxonomy. The new taxonomy was launched mid-September 2021.

This paper addresses and evaluates the methods used in the development of the new taxonomy. The methods include an expert review, mapping the practices at other statistical institutes, workshops with the participation of colleagues from the statistics-producing divisions in Statistics Denmark and, last but not at all least, several user tests.

The most fundamental change carried out by the implementation of the new taxonomy is a shift from a more traditional approach reflecting the organisation of the statistical production to an approach mainly based on the users' perspective. The new taxonomy has been rolled out on all our dissemination platforms as it is a vital principle, that the taxonomy is identical and familiar to our users across all platforms.

Based on these principles, the paper takes the discussion further and considers the possibility of creating a uniform topic taxonomy that can be implemented across all national statistical institutes, and elaborates whether this is a suitable option or not.

Keywords: topic taxonony, user perspective

Session 3.3. Metadata, reporting processes

When: Tuesday, August 23 at 15:30 - 16:25

Download presentation

Charlotte Nielsen, Senior Adviser, Methodology and Analysis, Statistics Denmark, cln@dst.dk

Rasmus Anker Stavad, Senior Adviser and Metadata Coordinator, Methodology and Analysis, Statistics Denmark, rkr@dst.dk

In recent years, Statistics Denmark has worked on implementing consistent and coherent metadata to support metadata driven production. The implementation follows international standards and focuses on increasing efficiency, by using standardized metadata actively and systematically in the production of official statistics and for dissemination to users. The content and organization of structural metadata follow the recommendations and standard outlined in UNECE's Generic Statistical Information Model (GSIM). The GSIM model excels by supporting the information objects that go into and out of the individual process steps of the Generic Statistical Business Process Model (GSBPM). This paper will present Statistics Denmark’s new and improved documentation portal, which showcase relationships between Quality reports (SIMS), classifications and code lists, as well as the documentation of registers and variables. Examples of coherent metadata and how they are presented for the users will be given.

Keywords: Coherent metadata, documentation portal, GSIM, GSBPM, standardisation, metadatadriven production

Session 3.3. Metadata, reporting processes

When: Tuesday, August 23 at 15:30 - 16:25

Download presentation

Jacob Chr. Solander, Web & StatBank, jcs@dst.dk

This paper describes how Statistics Denmark are using automated processes to ensure optimal workflow and minimum of difficulties and problems for the statistical offices sending data to Eurostat.

Like the other EU member states, Denmark and Statistics Denmark are obliged to send data series to Eurostat in the SDMX format. Statistics Denmark has always emphasized that sending data or reports to Eurostat must not be difficult or problematic for the individual statistical offices. That is why we have always worked to ensure that there is a single source for our data. That the statistical offices only need to update their data to one single place - the Sum-Database. From there, we have developed several fully automated processes that send data to Eurostat in the desired format using our existing disseminations tools and the StatBank.dk. The StatBank API forms the basis for the data transfer from Statistics Denmark and therefore this will be the key to future development areas.

Some of the points that presentation will review are:

• Automatically generated SDMX files to Eurostat

• Saved queries from the StatBank.dk to the IMF converted into SDMX format

• Queries for CENSUS 2021

• Use of API for the automatic processes

• Handling the transfer of large data sets

Keywords: Eurostat, International repporting, online communication, IMF, CENSUS 2021

Session 3.3. Metadata, reporting processes

When: Tuesday, August 23 at 15:30 - 16:25

Download presentation

Jan Bruusgaard, Statistics Norway, jbr@ssb.no

Most Nordic NSIs publish their statistics using a common statistical output database ("Statbank") interface: PxWeb and PxWebApi. Using this platform advanced users can access our data and statistics in a common way.

I will give an overview or status of the available tools that our external users and the NSIs themselves can use to extract data via the programming interface (API). Via API they can get updated figures into other applications like R, SAS, Excel, Business Intelligence software and websites. There are made programming libraries for our APIs which makes access, extraction and processing easier. Their code is published on sites for code sharing, like Github and CRAN. Such open source tools may furthermore prevent others from reinventing the wheel. Statistics Norway has done training, user support and marketing on our APIs in order to get users to extract data and utilize these API-tools.

The process on using API for extraction and the tools for automated publishing of updated figures can be illustrated by some cases. In the end I will show examples on free and paid products based on Statistics Norway’s statistics. These users retrieve data via API and often combine them with other data in a way we don’t do ourselves, eg https://telleren.kbnn.no. This kind of products are likely to grow in the future. The NSIs themselves should be aware of them and we could learn from them.

Keywords: Opendata, API, PxWebApi, dissemination, data-on-the-fly

Annex to paper: https://www.ssb.no/en/api/api-eksempler-pa-kode/_/attachment/download/0d5c33ca-f207-4257-a2f6-3da5d09b2bc9:b52edaae0fcd72045fcd3862a60a8ee5fcfa23d7/api-tools.xlsx

Session 3.4. Open data

When: Wednesday, August 24 at 10:30 - 11:25

Download presentation

Marie Haldorson, Statistics Sweden, marie.haldorson@scb.se

Providing official statistics as open data has been a key priority for Statistics Sweden for some years, in line with national and European open data strategies. Statistics Sweden initiated this work in 2017 by putting official statistics in the context of open data, with a user-centric approach guided by international examples. Since then, the ambition has been to provide official statistics as open data “that are of high quality, are well documented, respect data privacy concerns, are free and are easily accessible and usable” (UN seminar 2017).

An inventory at Statistics Sweden in 2018 showed which statistics was already provided as open data (through our statistical database) and which statistics had a potential as open data, typically standardized commissioned services with a lot of subscribers. “Raw data” or microdata was considered out of scope. Based on the inventory, funds were re-allocated to enable adding several tables to the database 2020, primarily local-level statistics on population by sex, age, income, education etc. These new tables also introduced a new local-level geography called DeSO (Demographic Statistical Areas), a division of the 290 municipalities into 5 984 smaller areas. Privacy concerns and the possibility to combine these new datasets with other open data guided the level of detail of the statistical content.

Other actions in 2020-2021 have been to automate DCAT-AP metadata from the statistical database so that all datasets are accessible from the national Data Portal and to differentiate the licensing of our open data. This enables flexible use of Statistics Sweden’s open data by Wikimedia and others, so that official statistics is easier to find on Google. The next goal is to make official statistics from the whole National Statistical System available as open data on the national Data Portal, enabling increased use and reuse.

Keywords: Open data, Local-level statistics

Session 3.4. Open data

When: Wednesday, August 24 at 10:30 - 11:25

Download presentation

Lars Pedersen, Statistics Greenland, larp@stat.gl

Junni L. Zhang, National School of Development, Peking University

John Bryant, Bayesian Demography Limited

To secure Democracy, the mission of Statistics Greenland as a National Statistical Institute (NSI) is to provide true and fair, transparent and comparable knowledge on our society.

Our primary dissemination tool is “Statbank Greenland”, where knowledge can be found as dimensional macro data, free and open, as in all Nordic Statbanks. All tables are reproducible, also programmatic. The Statbank systems have offered a golden opportunity for improved knowledge sharing, across NSI’s, researchers and the general public.

To bring knowledge sharing to a higher level, tables not only need to be detailed, but also need to be more informative and standardized, where possible. A special feature of “Statbank Greenland” that illustrates this point is in the area of Population Statistics.

Population forecasts, for instance, are important future predictors and decisive policy tools, where results stand on shoulders of true and fair statistics. In the case of component based population forecasts, detailed tables on the components are required.

A population account explains changes in population size from one point in time to the next by changes in birth- , death-, and migration components. An essential equality is that annual changes in births, deaths and migrations must add up to the change in the overall population size. A population account is the demographic equivalent of a National Account in economics. Statistics Greenland publishes such tables broken down by sex, age, cohort as Lexis-triangles. The tables allow users to fully reproduce population forecasts and any traditional demographic indicator, on both national and sub national levels.

A paramount feature of the Population Account is to highlight quality issues in the statistics. A correction post is needed to ensure equity. Analysis of corrections pinpoints possible trouble zones in the statistics. From Greenlandic experience the lesson learned is: if it smells fishy, it is fish.

If similar population account tables were published by multiple NSI’s, the transparency of Population Accounts would allow quality to be assessed and demographic details to be shared, paving way to a truly world class demographic resource for cross country comparisons and knowledge sharing.

Keywords: population, demography, knowledge sharing, transparency, quality

Session 3.4. Open data

When: Wednesday, August 24 at 10:30 - 11:25

Download presentation