Prof Mikael Jern 2014
Introduction to InfoVis and Geovisual Analytics Prof Mikael Jern
NCVA, Linköping University
”Discovery consists of seeing what everybody has seen and thinking what nobody has thought”
Albert von Szent-Gyorgyi (1893-1986)
http://ncva.itn.liu.se/
LET’S GO BACK TO THE VERY FIRST HISTORICAL KNOWN EXAMPLES OF INFORMATION
VISUALIZATIONS
Visualization is one of the oldest
communication media
Prof Mikael Jern 2014
Interest of the National Debt
Minard 1858 vs. EU NUTS2 2012
Cattle sent to Paris Age groups 0-15 and 65+
Prof Mikael Jern 2014
Information Visualization - Cholera outbreak London 1854
infected water pump?
Dr. John Snow:
Investigation of deaths from cholera
London, September 1854
death locations
spatial cluster
Information Visualization - Cholera outbreak London 1854
Prof Mikael Jern 2014
A good data representation is the key to solving the problem Information Visualization - Today
Industry
number of employees Point colour =
“förändring Rörelse”
Shape Colour =
“Företagsklimat”
http://ncomva.se/apps/dual/app/en/#story=0
The most famous example of an early Information Visualization!!
Minard’s graph from 1861 of Napoleon's march through Russia 1812
Prof Mikael Jern 2014
Geographic location in (X,Y) Flow map - Direction of movements Size of Army in flow (weighted arrows)
Temperature TIME!
Poland
temperature 422,000
100,000
4,000
60,000
Berezina River
The most famous example of an early Information Visualization!!
Napoleon's march through Russia 1812 from Poland to Moscow
Geographic location in (X,Y)
Flow map - Direction of movements Temperature temperature
Minard’s Graphics produced today with InfoVis!!
Napoleon's march through Russia 1812 from Poland to Moscow
Prof Mikael Jern 2014
AnotherMinard’s Graphics
Flow Map Visualization of French wine exports around 1864
World Trading - Collaboration between OECD and NCVA Trading with focus Japan 2009
Flow Map Visualization in InfoVis Today
Prof Mikael Jern 2014
Using Bar Chart AND Flow Map gives more Information (Knowledge)
Flow Map Visualization in InfoVis Today
http://mitweb.itn.liu.se/GAV/flowmaptrade/
Migration to and from Norrköping Kommun
Flow Map Visualization in InfoVis Today
http://mitweb.itn.liu.se/GAV/flowmapsweden/#
Prof Mikael Jern 2014
Sankey ”energy” chart 2013
http://www.iea.org/Sankey/
From Computer Graphics to Information Visualization ...
and Geovisual Analytics ...to InfoGraphics
Computer Graphics (vector drawings) Raster Graphics (pixel oriented) 1970:
1978:
1980:
1985:
1995:
Data Visualization Scientific Visualization Information Visualization
Prof Mikael Jern 2014
Example of InfoGraphics – This is NOT InfoVis!
http://www.linkedin.com/news?viewArticle=&articleID=904262641&gid=80552&type=member&item=79869 741&articleURL=http%3A%2F%2Fwww%2Eguardian%2Eco%2Euk%2Fworld%2Finteractive%2F2011
%2Fmar%2F22%2Fmiddle-east-protest-interactive-
timeline&urlhash=uwzt&goback=%2Egde_80552_member_79869741
Example of InfoGraphics – This is NOT InfoVis!
Prof Mikael Jern 2014
Example of Information Visualization (Visual Analytics)
Volume Rendering
Isosurfaces
SciVis
Physical data
Glyphs
Scatter Matrix Table Lens
Parallel Coordinates
InfoVis
Abstract data
Prof Mikael Jern 2014
SciVis = Physical Data (human body, earth, molecules, physical space
InfoVis = Abstract Data (statistical, financial, business information, text documents
Prof Mikael Jern 2014
3D InfoVis vs. 2D InfoVis
from one 3D view to multiple linked views
3D InfoVis vs. 2D InfoVis
from one 3D view to multiple linked views
Prof Mikael Jern 2014 TIME
Country
Immigrants
Multiple Time Shaded “3D Curves”
Prof Mikael Jern 2014
Energy
Consumption
3D Scatter Plots were popular in the late 80s
Scatter Plot – simple
Prof Mikael Jern 2014
Why Information Visualization?
Massive statistical and business information … and is growing
Which information is important?
Gain insight and knowledge
“a picture is worth a thousand words”
•Domestic sales larger than International and growing;
•Flat international sales and decreases sharply in August;
•Cyclical sales pattern in Domestic sales repeated on a quarterly basis reaching a peak in last month of quarter;
What these numbers could not communicate when presented as text in a table, which our brains interpret through the use of verbal processing, becomes visible and understandable when communicated visually. This is the power of “statistics data visualization."
Cyclical sales
Why Information Visualization?
Prof Mikael Jern 2014
Search + Examine + Explain =
SEE
to search for meaningful patterns … Discovery
then examine them and once they’re found…. Gain Understanding
to communicate meaningful findings to others ….. Explain
“aha, I see!”
Why Information Visualization?
Instead of One 3D View used in SciVis InfoVis apply multiple linked Views - “Dashboards”
Why Information Visualization?
Prof Mikael Jern 2014
………InfoVis now also on mobile devices using dashboards
(requires HTML5/JS)
Why Information Visualization?
http://mitweb.itn.liu.se/GAV/dashboard/
………InfoVis now also on mobile devices using dashboards
(requires HTML5/JS)
Why Information Visualization?
http://epp.eurostat.ec.europa.eu/cache/RSI/
Prof Mikael Jern 2014
World Statistical Data
1. ASK A SPECIFIC QUESTION
Where do we have high Fertility Rate?
2. GATHER YOUR INFORMATION Get data from World Databank….
3. VISUALLY REPRESENT THE MENTAL MODEL Select most suitable Visualization Method
Information Visualization in 4 easy steps
Prof Mikael Jern 2014
4
.
RESULT IN A SCATTER PLOTSize: Population
Time: 1960, 1961, 1962,..
Colour: Fertility Rate
Information Visualization in 4 easy steps
(X,Y): Age 0-14 vs. Fertility Rate
Age 0-14
Fertility
WORLD STATISTICAL DATA
Prof Mikael Jern 2014
Information Visualization Definition: Two Mantras
Overview, zoom & filter, details-on-demand (Shneiderman - 1996)
Analyze first, show the important, zoom, filter and analyze further and details-on-demand (Keim - 2006)
1. OVERVIEW - SCAN THE BIG PICTURE
2. ZOOM & FILTER - SEARCH SPECIFICS
3. DETAILS ON DEMAND - LINK TO FURTHER DETAILS
Prof Mikael Jern 2014
1. OVERVIEW - SCAN THE BIG PICTURE
2. ZOOM & FILTER - SEARCH SPECIFICS
3. DETAILS ON DEMAND - LINK TO FURTHER DETAILS
1. OVERVIEW - SCAN THE BIG PICTURE
2. ZOOM & FILTER - SEARCH SPECIFICS
3. DETAILS ON DEMAND - LINK TO FURTHER DETAILS
Prof Mikael Jern 2014
1. OVERVIEW - SCAN THE BIG PICTURE
2. ZOOM & FILTER - SEARCH SPECIFICS
3. DETAILS ON DEMAND - LINK TO FURTHER DETAILS
http://mitweb.itn.liu.se/GAV/dashboard/#story=data/Ageing Population in the World.xml&layout=[map,(scatterplot,barchart)]
Hans Rosling’s world
Prof Mikael Jern 2014
Scatter Plot – how many (attributes)?
World Countries
Population ages 0-14 vs. Life expectancy at birth;
Colour: Population ages 65+ ; Circle Size: Total population; Trails: Time;
Life expectancy
age 0-14
age 65+
Size: Population Time: 1960, 1961,..
Colour: age group 65+
“Scatter Matrix”
InfoVis method
Prof Mikael Jern 2014
fertility
fertility rate
“An Interactive visualization is worth a thousand pictures”
Spatial – Time - Variable
http://mitweb.itn.liu.se/GAV/dashboard/#story=data/Ageing Population in the World.xml&layout=[map,scatterplot]
Prof Mikael Jern 2014
Information visualization is the use of interactive, visual representations of abstract data (but as you have seen often with a spatial dimension) and to use perception to amplify cognition.
It is the process of forming a mental model of data, thereby supporting insight into that data.
Information Visualization Definition
…….forming a mental model of data, thereby supporting insight into that data….
1960 1985
2009
Prof Mikael Jern 2014
forming a mental model of data, thereby supporting insight into that data – Sweden municipalities age group 0-14 (1960-2010)
use COLOUR
time
1960 1985 2010
…….forming a mental model of data, thereby supporting insight into that data…. “use lines for time movements”
Prof Mikael Jern 2014
Information visualization is the use of interactive, visual representations ... and to use perception…
Importance of Human User Interface
Find the red square?
Information Visualization helps users …
Find patterns, outliers and trends
Prof Mikael Jern 2014
Find the blue circle?
Information Visualization helps users …
Find patterns, outliers and trends
"The holy grail of
information visualization is to make the insights stand out from otherwise chaotic and noisy data."
Information Visualization helps users …
Find patterns, outliers and trends
Net Migration
Prof Mikael Jern 2014
Net Migration in divided colour vs. grey scale
Net Migration Net Migration
Information visualization is the use of interactive, visual representations of abstract data (but as you have seen often with a spatial dimension) and to use perception to amplify cognition.
It is the process of forming a mental model of data, thereby supporting insight into that data.
Information visualization helps users:
Find patterns, outliers and trends
Explore data to build intuition, understanding and knowledge
Communicate understandings and knowledge to others
Information Visualization Definition
Prof Mikael Jern 2014
The Perfect Car Data Set applied to InfoVis
“Visualization of multivariate abstract data”
Abstract Data in Information Visualization
Multivariate - Quantitative data and Categorical data
Data Items
Prof Mikael Jern 2014 Miles Per Gallon
Acceleration 0-60
Size: Weight Color: Horsepower
Scatter Plot – with high correlation
http://mitweb.itn.liu.se/GAV/mdim/#story=2
Weight
Acceleration 0-60
Size: Price Colour: Miles per gallon
Scatter Plot – with high correlation
Prof Mikael Jern 2014
Scatter Plot – how many (attributes)?
1D: Weight vs. 2D: Acceleration 0-100 Circle Size: Price;
Colour: Miles Per Gallon;
Prof Mikael Jern 2014
Multivariate - Quantitative data and Categorical data
Data Items
Categorical Quantitative Categorical (Ordinal) Quantitative
Country
Acceleration 0-60
Prof Mikael Jern 2014
The Perfect Car Data Set applied to InfoVis
“Visualization of multivariate abstract data”
http://mitweb.itn.liu.se/GAV/mdim/#story=2
Multidimensional Visualization of Multivariate Data
Information Visualization Definition
Prof Mikael Jern 2014
Country
Indicators Time
Italy; Fertility; 2010
Italy; [indicators]; [time]
2010 1960 Fertility GDP Population
Fertility;[data item]; [time]
2010 1960
Spatio-Temporal and Multivariate Visualization using a Data Cube
Spatio-Temporal and Multivariate Visualization using a Data Cube
Prof Mikael Jern 2014
Introducing ”Treemap” – using hierarchical data
Rectangle Size = Population Colour = Fertility rate
Hierarchy = Continent-Country
Göteborg Stad tidig med Publicera Öppna Data
• Statistik visas på ett sätt som engagerar många i stadens utveckling;
• Trender över tid ökar förståelsen;
• Statiska rapporter ersätts med interaktiva webb sidor
Prof Mikael Jern 2014
http://epp.eurostat.ec.europa.eu/cache/RSI/
http://ncva.itn.liu.se/great-statistics-visualization/eurostat-statistics-visualization?l=en
http://epp.eurostat.ec.europa.eu/cache/RSI/
http://ncva.itn.liu.se/great-statistics-visualization/eurostat-statistics-visualization?l=en
Prof Mikael Jern 2014
SCB Statistik Atlas
• Forskningssamarbete sedan 2004;
• Statistikatlas 2010;
• Storytelling;
• Visuella Nyhets- bulletiner från 2012;
• Många efterföljare;
http://www.scb.se/kartor/Statistikatlas/index.html
SCB Statistik Atlas
Nya Data
Prof Mikael Jern 2014
• Ett urval av SCB’s egna indikatorer
• Över tid 1968-2011
• Integrerad Storytelling
http://www.scb.se/kartor/Statistikatlas/index.htm l
SCB Statistik Atlas
Västra Götalandsregionen
http://www.vgregion.se/sv/Vastra-Gotalandsregionen/startsida/Regionutveckling/Publikationer-statistik/Interaktiv-statistik1/
Prof Mikael Jern 2014
Västra Götaland eXplorer Arbetslöshet
http://www.vgregion.se/sv/Vastra-Gotalandsregionen/startsida/Regionutveckling/Publikationer-statistik/Fakta-och-statistik/Interaktiv-statistik/
Arbetslöshet Totalt
Arbetslöshet Kvinnor
Arbetslöshet Män
Arbetslöshet Totalt
Story Arbetslöshet 1992-2011
Prof Mikael Jern 2014
Graphical Excellence requires Good Perception
Graphics excellence
….is the well-designed presentation of
interesting data, a matter of substance, statistics, and design
……consists of complex ideas communicated with: clarity, precision, and efficiency
……is that what gives the viewer: the greatest number of ideas, in the shortest time, with the least ink, in the smallest space
……is nearly always multivariate
…..requires telling the truth about the data
Tufte
http://nvac.pnl.gov/
Introducing GeoVisual Analytics an extension to InfoVis ....in 2005
Prof Mikael Jern 2014
Geographic Visualization
Geovisual Analytics
Communicate Storytelling
Publish
Data Multiple Sources
Dynamic Filter Information Visualization
Cognitive Perceptual
Science
Time Animation Visualization
http://ncva.itn.liu.se/?l=en
Interactive Visualization is nice to play with but…
Difficult to collect and report the results knowledge gained in an Explorative Analytics session ;
Use technologies that enable analysts to communicate what they know through use of appropriate visual
metaphor and principles of reasoning and graphics Challenge and Motivation for Geovisual Analytics ..
Prof Mikael Jern 2014
Big Data - http://ncva.itn.liu.se/big-data-geovisual-analytics?l=en 10,000 geographical regions
The science of analytical reasoning facilitated by interactive visual interfaces – e.g. dynamic linked multiple views;
Exploring and analyzing spatial-temporal and multivariate data;
Discern trends or patterns - derive insight and draw conclusions;
Communicate discovery and knowledge effectively for action
with 100% Web compliant;
Moving Research into Practice;
Introducing GeoVisual Analytics an extension to InfoVis ....
Prof Mikael Jern 2014
Gather Data and Information – Tasks?
Visual representation
Choose layout and visual forms that aid analysis
Develop insight – Interactive session Through exploration – What is important?- Filtering – Tell the Story
Produce results (knowledge) Storytelling, Presentation, Collaboration and Publishing
Visual Analytics Reasoning Process – Sense Making Loop
Sense-making is an effort to understanding, to recognize particular
characteristics of the data and understand what they mean .. “it make sense”
Motala River – Concentration of Nitrogen
“many cows along the river”
Prof Mikael Jern 2014
Ericsson Research
Visualization of Self-Organizing Networks Operated by
Automatic Neighbour Relations• Growth of cellular radio networks
• Need of automatic algorithms
• Automatic Neighbour Relations (ANR) developed by Ericsson
• ANR must be proven to gain network operators’ trust
Should we trust automatic algorithms?
(Visualization of Self-Organizing Network)
VoSON
Prof Mikael Jern 2014
CGI 657
PCI 481 CGI 671
PCI 11
If the PCI is unknown, the CGI is requested.
The ANR Procedure
Ericsson Research – VINNOVA
Search for mobile cell anomalies
Prof Mikael Jern 2014
Ericsson Research – VINNOVA
Search for mobile cell anomalies
Visual Analytics at Ericsson - Mobile accessibility
http://mitweb.itn.liu.se/GAV/anross.wmv http://ncva.itn.liu.se/anross-mobile?l=en
Prof Mikael Jern 2014
Dashboard multiple views application
http://mitweb.itn.liu.se/GAV/dashboard/
…….. now also on mobile devices using dashboard
Prof Mikael Jern 2014
http://ncva.itn.liu.se