Information Visualization Robert Spence Imperial College London What is it?

We convert data into a picture We look at that picture and gain insight

Many people and institutions possess data that may ‘hide’ fundamental relations Realtors Bankers Air Traffic Controllers Fraud investigators Engineers Baseball fans They want to be able to view some graphical representation of that data, maybe interact with it, and then be able to say Who is looking at data?

.... Now that is interesting We are not discussing statistical analysis, which might - or might not - come later. We are concerned solely with the Ah HA ! Acquisition of insight

The Serious Fraud Office (SFO) suspected mortgage fraud The SFO provided 12 filing cabinets of data After 12 person years a suspect was identified The suspect was arrested, tried and convicted The data was supplied in electronic form A visualization tool (Netmap) was used to examine the data After 4 person weeks the same suspect was identified A master criminal behind the fraud was also identified Example: Fraud Detection Is information visualization useful?

manufactures microprocessors on silicon wafers that are routed through 400 steps in many weeks. This process is monitored, gathering 140,000 pieces of information about each wafer. Somewhere in that heap of data can be warnings about things going wrong. Detect a bug early before bad chips are made. has 1500 scientists using an advanced information visualization tool (Spotfire) for decision making. “With its ability to represent multiple sources of information and interactively change your view, it’s helpful for homing in on specific molecules and deciding whether we should be doing further testing on them” Drugs and Chips Fortune Eli Lilly Sheldon Ort of Eli Lilly, speaking to Fortune Texas Instruments Is information visualization useful?

Visualize: to form a mental model or mental image of something Visualization is a human cognitive activity, not something that a computer does A definition:

The formation and inspection of a mental map of the London Underground system

M. Minard’s map of Napoleon’s march to, and retreat from, Moscow A record of data Minard incorporated many variables: Line width = size of army Latitude and longitude identified Colour identifies the direction of the march The temperature is indicated

The cholera epidemic, London 1845 Dr. John Snow, medical officer for London, investigated the cholera epidemic of 1845 in Soho. He noted that the deaths, indicated by points, tended to occur near the Broad Street pump. Closure of the pump coincided with a reduction in cholera.

On 28th January 1986 the space shuttle Challenger exploded, and seven astronauts died, because two rubber O-Rings leaked. The Challenger disaster The previous day, engineers who designed the rocket opposed the launch, concerned that the O-Rings would not seal at the forecast temperature (25 to 29 o F). After much discussion, the decision was taken to go ahead. Cause of the accident: An inability to assess the link between cool temperature and O-Ring damage on earlier flights. Many charts poorly presented

A scatterplot showing the experience of all launches prior to the Challenger, revealing the serious risk of a launch at 29 o F Temperature forecast The Challenger disaster After Tufte

It facilitates interaction, and rearrangement of a display “A graphic is no longer ‘drawn’ once and for all: it is ‘constructed’ and reconstructed (manipulated) until all the relationships which lie within it have been perceived... A graphic is never an end in itself: it is a moment in the process of decision making” Bertin, 1981 How can the computer help?

Data about the success (black) and failure (white) of applying different treatments(A to G) to a range of crops (1 to 10)

Communication data: records of telephone calls Rearrangement: Example 1 It is difficult to gain any insight from this table

Rearrangement as a node-link graph can help Rearrangement: Example 1

More rearrangement really does help! Rearrangement: Example 1 Exactly the same graph - just pulled apart!

Cleveland, 1985 “Graphing data needs to be iterative because we often do not know what to expect of the data: a graph can help discover unknown aspects of the data, and once the unknown is known, we frequently find ourselves formulating new questions about the data.” The power of rearrangement

Proust “The real voyage of discovery consists not in seeking new landscapes but in having new eyes”

Representation Presentation Interaction/Navigation Human performance

We have seen examples of visualization - but what are the issues?