The Bactra Review Atlas of the
European Novel
A mis-placed affection. It's not impossible to get useful information from a
time-series; it's even easier than making a silk purse from a sow's ear,
usually. It helps if you've got a nice fast computer with nothing better to
do, heaps of high-quality data, and a pretty heavy mathematics. This is why
time-series analysis is a very active area of research, and the Handbook
of Econometrics is four volumes on the order of a thousand pages a
piece. More to the present point, it's why Edward Tufte never tires of
anathemizing time-series in his books on the visual display of quantitative
information. Those who want to learn about causal relations (like Moretti, who
wants to understand how social factors affect literary forms) would have to
work hard to find a worse way of representing their data. (So of
course I'm researching the best way of squeezing causal models from
time-series: like blood from stones...)