Withdraw data and put it into a table, then write a brief summary.

Tufte measured what he calls data density (numbers per square inch) in the statistical graphics in various science journals. He found, for instance, that Nature’s tables showed a minimum of 3 numbers per square inch (not good), a maximum of 362 (much better), and a median of 48 (pretty good, as a median). Science, however, had a minimum of 5, and a maximum of 44, with a median of 21—not nearly as dense as Tufte would approve. The New England Journal of Medicine had a minimum of 3, and a magnificent maximum of 923, but fell to a median of 12—a severe disappointment. Two newspapers, and one business journal outdid the medical journal. The Wall Street Journal had the best median of this group, with 19 numbers per square inch; both Fortune and the London Times came close behind, with 18. Interestingly, the worst newspaper on Tufte’s list, the old Commie outlet Pravda, had a maximum of one number per square inch, a minimum of 1/10th of a number per square inch, for a median of 0.2 numbers per square inch. Overinflated. So there is often a wide variation within each publication, but overall, "the average published graphic is rather thin."

Answer