Reading a Picture

Visual storytelling

Visualising data helps understanding facts.
Sometimes it’s very easy to understand a graph; sometimes it’s necessary to read it and to study it to discover unknown territory.

Such graphs are little masterpieces. Here’s one of these and I am sure the authors had more than one iteration and discussion while creating it.
The graph tells the story of the average disposable income and savings of households in Switzerland, published by the Swiss Federal Statistical Office FSO.

snip_disposable-income2

The authors kindly give a short explanation:

How to read this graph.
In one-person households aged 64 or under, the upper-income group has a disposable income of CHF 8487 per month and savings of CHF 2758 per month. Representing 4.0% of all households, this income group corresponds to a fifth of one-person households aged 64 or under (20.1%)

There’s another nice graph, a little bit less elaborated, also explained by the authors:

snip-povertyrates

Statistics ♥

But there’s one thing that is not explained:

snip_poverty-cithe confidence interval!

‘A confidence interval gives an estimated range of values which is likely to include an unknown population parameter, the estimated range being calculated from a given set of sample data,‘ and the above poverty data are from a sample of ‘approximately 7000 households, i.e. more than 17,000 persons who are randomly selected…’.
Or:
The confidence intervals for the mean give us a range of values around the mean where we expect the “true” (population) mean is located (with a given level of certainty, see also Elementary Concepts). ….. as we all know from the weather forecast, the more “vague” the prediction (i.e., wider the confidence interval), the more likely it will materialize. Note that the width of the confidence interval depends on the sample size and on the variation of data values…..’

Khan Academy gives lectures about topics like confidence intervals, sampling, etc.

snip_20161129160845.

Which one ?

The above graphs use just one of multiple possibilities for visualising data.

snip_graph-catalogue

Severino Ribecca’s Data Visualisation Catalogue is one of many websites trying to give an overview. And there’s the risk to get lost in these compilations.

snip_swimring                            © listverse.com

LOD MOOC

Massive Open Online Courses (MOOC) are available worldwide and offer tons of topics, also about Linked Open Data (LOD). An easy way to enter the semantic web.

Two examples:

HPI

snip_opehpi

The Hasso Plattner Institute, Potsdam provides, for some years now, a course in Linked Data Engineering with a certificate. I did it some years ago and enjoyed it.

FUN (INRIA)

snip_lodcourse2

The French platform FUN offers a LOD course, too. (Thanks to Adrian at zazuko.com for the hint)

And books

Step by step Bob DuCharme introduces RDF, SPARQL, LOD …

snip_ducharmesparql

.

snip_ducharmesparql-preface

IMAODBC 2016: And the winner is…

The Best Presentation Award of the International Marketing and Output Database Conference IMAODBC 2016 in Gozd Martuljek, Slovenia goes to Susanne Hagenkort-Rieger and her team from DESTATIS (Statistisches Bundesamt, Germany).

In her presentation Susanne highlighted the importance of web search statistics  and why intuition when emphasizing selected statistical data is often not sufficient. To achieve relevance and accessibility of most popular statistical data we should not ignore what the web search data say.

a12fd9e2e59120fadd5cfb91863759d1

Presentation is available at the IMAODBC 2016 website…

A few facts about IMAODBC 2016 as presented in the second best presentation by Corey Jenkins, USDA – Foreign Agricultural Service, U.S.A.:

IMAODBC-2016-ig.jpg

 

 

20 Years Ago

1996

On the 2nd of September 1996, Statistics Switzerland published its brand-new website, www.bfs.admin.ch. It was one of the first (if not the first) of the Swiss Administration (www.admin.ch).

info-internet

In three languages…

… and already with quite rich structure and content.

SwissStats-April1997

The Wayback Machine …

… shows the developments since 1996

snip_waybackmachine

https://archive.org/web/.

1996:
Handmade with Frontpage as editing software

SwissStats-fields

https://web.archive.org/web/19970502093430/http://www.admin.ch/bfs/stat_ch/eber_m.htm

November 2004:
New layout, made with Day Communiqué as Content Management System and a database for file download

StatSchweiz-November2004

December 2007 ff
Layout adapted to the general layout of admin.ch. The same CMS now bought by Adobe and renamed Adobe Experience Manager AEM

StatSchweiz-Dezember2007

snip_bfs-2016

The next one …

… must be based on a new technology:

  • CMS remains state of the art for content presentation
  • Assets come from databases
  • Web services (via a web service platform) deliver assets from databases to the presentation platform.

And with such a three-layer architecture the website will be able to display data from ubiquitous databases and also offering data to third parties via web services: Open data compatible.

Disrupting Dissemination – From Print to Bots

Digitally disrupted data production

“The collection of statistics has been digitally disrupted, along with everything else, and there are important questions about collection methods and whether or not “big data” genuinely offers promise for a giant leap forward in the productivity of official statistics.”

This statement in Financial Times’ edition of August 20th, 2015 deals with UK’s Office for National Statistics ONS. Its title:  “UK needs a statistical strategy to catch up with digital disruption”. Its message: ONS (and I think all Official Statistics) has problems to keep up “with the profound changes in the structure of the economy during recent decades.”

The “Independent Review of UK Economic Statistics” by Professor Sir Charles Bean, Professor of Economics at the London School of Economics in March 2016 goes deeper and gives 24 recommendations, some of these obviously valid for statistics’ production and producers in general.

snip_Beans-independentReview

“Innovation and technological change are the wellspring of economic advancement. The rapid and sustained rise in computing power, the digitisation of information and increased connectivity have together radically altered the way people conduct their lives today, both at work and play. These advances have also made possible new ways of exchanging goods and services, prompted the creation of new and disruptive business models, and made the location of economic activity more nebulous. This has generated a whole new range of challenges in measuring the economy.” (p.71)

“Measuring the economy has become even more challenging in recent times, in part as a consequence of the digital revolution. Quality improvements and product innovation have been especially rapid in the field of information technology. Not only are such quality improvements themselves difficult to measure, but they have also made possible completely new ways of exchanging and providing services. Disruptive business models, such as those of Spotify, Amazon Marketplace and Airbnb, are often not well-captured by established statistical methods, while the increased opportunities enabled by online connectivity and access to information provided through the internet have muddied the boundary between work and home production. Moreover, while measuring physical capital – machinery and structures – is hard enough, in the modern economy, intangible and unobservable knowledge-based assets have become increasingly important. Finally, businesses such as Google operate across national boundaries in ways that can render it difficult to allocate value added to particular countries in a meaningful fashion. Measuring the economy has never been harder.” (p. 3)

And: “Recommended Action 4: In conjunction with suitable partners in academia and the user community, ONS should establish a new centre of excellence for the analysis of emerging and future issues in measuring the modern economy.”  (p.118)

Disrupting Dissemination of Statistics

The rise of new technologies followed by new information behavior has also disrupted existing dissemination formats (from print to digital) and dissemination practices (from quasi-monopolistic to open and multiple).

A well-known example for disrupting dissemination is given by Wikipedia and its subject is Wikipedia itself:
“The free, online encyclopedia Wikipedia was a disruptive innovation that had a major impact on both the traditional, for-profit printed paper encyclopedia market (e.g., Encyclopedia Britannica) and the for-profit digital encyclopedia market (e.g.,Encarta). The English Wikipedia provides over 5 million articles for free; in contrast, a $1,000 set of Britannica volumes had 120,000 articles.” (Article: https://en.wikipedia.org/wiki/Disruptive_innovation)

In fact, disruptive tendencies happen on both sides: in producing and in presenting or accessing statistical information.

Some thoughts about this:

  1. Until the end of the 20th-century, print was the main channel for disseminating statistics. Libraries in Statistical Offices and Society had their very vital role.
  2. With the internet opened a new channel: Statistical Offices’ Websites appeared, access to databases and attractive data presentation (visual, storytelling, see i.e. this) were top themes and stuff for long discussions. Access to data was now simple and for everyone.
  3. With the open data initiatives not only accessing but also disseminating statistical information got much easier. Nearly everyone could become a data provider. License fees no longer hindered the redisseminaton of official statistics and APIs or webservices provided by statistical offices made this possible in an automated way.
    Statistics can be easily integrated into websites and apps of non-official data providers, this with all the chances to enable democratic conversation and the risks of data misuse.
  4. All this gives statistics a much more important role in communication processes. On the other hand communicating with statistics gets simpler: Letters, telephone calls and even e-mails become cumbersome seen the possibility bots (will) provide. With a stats bot in my daily used messenger, I ask for a statistical information, and the bot uses a search engine or connects me directly to a statistical expert.
    Brands that already have full-fledged apps and responsive websites can take advantage of bots’ ability to act as concierges, handling basic tasks and micro-interactions for users and then gracefully connecting users with apps or websites, as appropriate, for a more involved experience.” (Adam Fingerman, venturebeat, 20.7.2016)
  5. What’s next? Innovation with disruption goes on, but disruption does not always mean destruction: It’s still a wise decision to keep some information in paper format. A statistical yearbook with key data lasts for centuries, not so a website, an API or a bot.