• clear and concise
• encourages understanding

• allows us to assess cause and effect
• shows the data at multiple levels of detail

The aim is to induce the viewer to think about substance over methodology, graphic design and technology.

The graphics should reveal information (not obscure it).

A tenth century graph. Is this the first ever graph?

from H Gray Funkhauser, "A note on a tenth century graph", Osiris 1 (1936), 260-262, reproduced on p.28 of E R Tufte, The Visual Display of Quantitative Information, Cheshire, CT: Graphics Press 1983

Graphs use abstract, non-representational pictures to show numbers.

Statistical graphs were invented around 1750-1800.

Graphs allow us to reason about quantitative data. (They can also be misused.)

The space shuttle Challenger exploded, 73 seconds after launch, on January 28, 1986, killing the seven astronauts on board.

The shuttle failed because two rubber O-rings leaked fuel.

The O-rings lost their resiliency because the ambient temperature was very low.

Before the flight, engineers knew of the problems of O-ring damage with low ambient temperatures and recommended that the shuttle not be launched. So why did the flight go ahead?

Instead of documenting the history of failed O-rings using a simple and effective graph, the information was presented as it appears at left: as a sequence of little rocket drawings, with a complex legend of shading, and numbers scattered around on the rocket bodies indicating failed O-rings and launch temperature.

As an example, the information designer Tufte redrew this information in a much simpler form clarifying the relationship between launch temperature and O-ring failure and omitting irrelevant information such as the order in which the failures occurred. You can have a look at the result online.

London Underground map, designed by Harry Beck in 1933 achieves a reduction of complexity through:

• simplified geometry and topology
• a focus on connections rather than geographic accuracy

New York City Subway map (1972, 1977-1978, 1979 onwards).

• Too much reduction makes the representation too abstract and reduces understandability.

Of course mapping is a fantastic information display mode in many contexts.

• French Piste Map (provides a photographic style overview)
• UK Census map (excellent use of colour to show density)
• Snow's famous Cholera Epidemic map (a landmark geographical information system)
• A topographical map of mediocre quality but illustrative nevertheless. (What makes this map mediocre? How could it be improved?)

Original design used a single colour vector display.

Extensive colour research allowed designers to come up with a perceptually-based system to help ease fatigue and draw operators' attention to critical data.

Do a web search to find out what these displays look like. Here is good example.

How do the principles of interface design (presented at the end of the previous lecture) impact on the design of these displays?