Skip to content
Closed
Changes from all commits
Commits
File filter

Filter by extension

Filter by extension

Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
63 changes: 49 additions & 14 deletions 02_activities/assignments/assignment_2.md
Original file line number Diff line number Diff line change
Expand Up @@ -9,27 +9,62 @@
- You can find data visualizations at https://public.tableau.com/app/discover or https://datavizproject.com/, or anywhere else you like!
- For each visualization (good and bad):
- Explain (with reference to material covered up to date, along with readings and other scholarly sources, as needed) why you classified that visualization the way you did.
```
Your answer...







I classify Minard's flow map of Napoleon's 1812 Russian campaign as a strong
visualization for three reasons.

First, it achieves an exceptionally high data-ink ratio (Tufte, "The Visual
Display of Quantitative Information," 1983). Six variables — army size, two
spatial coordinates (latitude/longitude), direction of travel, location relative
to dates, and temperature on the retreat — are encoded without decorative
clutter. Nearly every mark carries information, which is the principle Tufte
uses to define graphical excellence.

Second, it uses pre-attentive visual channels effectively. The width of the
band redundantly encodes the magnitude of the army through position and size,
two of the most accurate perceptual channels in Cleveland and McGill's (1984)
ranking of elementary graphical-perception tasks. The collapse of the band from
~422,000 to ~10,000 men is legible instantly, before any reading of numbers
occurs.

Third, it tells a story (a stated learning outcome). The dark outbound flow and
the pale returning flow, anchored to a temperature line below, let the viewer
reconstruct cause and consequence — the catastrophic toll of the winter retreat —
in a single integrated frame. This narrative integration of space, time, and
magnitude is why Tufte called it possibly "the best statistical graphic ever
drawn."

I classify the Fox News ACA-enrollment bar chart as a poor visualization for three reasons.

First, it violates the proportional-ink / Lie Factor principle (Tufte 1983). The chart compares a March 27 enrollment figure of about 6 million to a target of 7 million, but the 6-million bar is drawn at roughly one-third the height of the 7-million bar. The bars' visual sizes are wildly disproportionate to the values they represent; Tufte's Lie Factor (size of effect shown ÷ size of effect in data) is far from the ideal of 1, manufacturing a sense of failure not present in the numbers.

Second, it manipulates the baseline. The y-axis does not begin at zero. For bar charts, where the encoded quantity is bar length/area, a non-zero baseline distorts the very channel the reader is meant to decode (Cleveland & McGill, 1984, identify length/position judgments as the perceptual basis of bar charts). Truncating that baseline breaks the mapping between data and ink.

Third, it fails the test of integrity and equity. Cairo (*The Functional Art*, 2013; *How Charts Lie*, 2019) argues a chart's first duty is to be truthful and to let readers draw their own conclusions. This graphic instead engineers a predetermined political conclusion, encoding rhetoric rather than data — the opposite of supporting critical, informed consumption.

```
- How could this data visualization have been improved?
```
Your answer...



Minard's flow map of Napoleon's 1812 Russian campaign:
First, the graphic relies on a brown/black versus tan colour contrast to
distinguish the advance from the retreat. This single redundant cue can fail for
viewers with low vision or in poor reproductions; adding a clear directional
label or arrowheads on each band, and testing the palette for sufficient
luminance contrast (WCAG-style guidance), would make the advance/retreat
distinction non-reliant on a single channel.

Second, the temperature scale at the bottom is in Réaumur degrees and the
distances/dates are sparse, demanding prior context most modern readers lack. A
brief integrated legend converting units (Réaumur to Celsius), and a short title
caption stating the question being answered, would lower the barrier to entry and
make the chart self-explanatory for a general audience rather than only for
readers already familiar with the campaign.



First, and most fundamentally, the y-axis should start at zero so that bar heights are proportional to the values (Tufte 1983; Cairo 2019). With a zero baseline, 6 million versus 7 million reads as a modest shortfall — roughly 86% of target — rather than a near-total collapse. This single change restores a Lie Factor close to 1 and lets the bars do honest perceptual work.

```
Second, the chart should add context and accurate labelling to tell the real story. Showing the enrollment trajectory over time (a line or column series across the sign-up period) rather than two isolated snapshots would convey momentum an the late surge, and clear value labels plus a neutral, descriptive title would let the audience interpret the data themselves. Choosing an appropriate chart type for a part-to-whole-against-target comparison — for example a labelled progress bar o bullet chart (Few, *Show Me the Numbers*, 2004) — would communicate "progress toward goal" far more truthfully than two mismatched columns.

- Word count should not exceed (as a maximum) 500 words for each visualization (i.e.
300 words for your good example and 500 for your bad example)

Expand Down
Loading