The Ilonka Reader

Notes on the Books I Have Read

Month: March, 2016

The Visual Display of Quantitative Information

By Edward Tufte.

I quickly read this book for the second time recently in order to take notes on it. My notes are below.

Data maps:

  • Dr John Snow’s cholera map of 1854 (pg 24)
  • Charles Joseph Minard’s map of French wine exports in 1864 (pg 25)
  • population distribution usa, 1970 (pg 156)

 

Time-series:

  • Don’t waste on simple data, use numbers for that. Use for highly variable data.
  • NYC weather for 1980 (pg 30)
  • E. J. Marey’s graphical train schedule for Paris to Lyon in the 1880s. (pg 31)
  • J. H. Lambert (early 1700s) and William Playfair (last 1700s) were the great inventors of modern graphical design. (pg 32)

 

Narrative graphic:

  • Minard’s fate of Napoleon’s army in Russia, drawn 1861 (pg 41)

 

Abstract design:

  • Thermal conductivity of copper, 1974 (pg 49)

 

Experiments have shown that we don’t perceive measured data accurately.

  • e.g. perceived area of circle grows more slowly than actual:
  • reported perceived = (actual area)^x, x=.8+-.3 (pg 55)
  • plus you can tell that everyone is different

 

Representations of 3d objects lead to confusing comparisons as perspective reduces our ability to accurately compare measurements.

 

  • “Representation of numbers should be directly proportional to the numerical quantities represented.”
  • “Show data variation, not design variation.” (pg 61)
  • “Number of … variables depicted should not exceed the number of dimensions in the data.” (pg 71)
  • “Graphics must not quote data out of context.” (pg 74)

 

No one can write decently who is distrustful of the reader’s intelligence, or whose attitude is patronizing. – E. B. White

 

  • Data-ink ratio = data-ink / total ink used to print graphic
  • erase non-data-ink and redundant data-ink
  • (it’s only a single dimension to measure with, efficiency, should also consider complexity, structure, density, and beauty. pg 137)

 

Chartjunk to erase:

  • unintentional optical art (patterned shading.”moire vibration”)
  • grid lines (can be faint)
  • pg 118: perhaps the worst graphic ever, only 5 pieces of data
  • the duck (decorative forms)

 

Tufte forms:

  • bar plot pg 128
  • boxplot pg 129
  • scatter plot pg 133

 

Data can be part of graphic, numbers as points or labels as markers for max/min or points values.

 

  • data density = numbers of entries in data matrix/area of data graphic
  • e.g. annual sunshine report in London 1971 (pg 165)

 

Graphs can often be shrunk.
Graphical elegance is often found in simplicity of design and complexity of data.” (pg 177)

Advertisements

What is a p-value anyway?

By Andrew Vickers.

Another saunter around statistics, this book is a humorous approach to understanding many statistical techniques used (both correctly and incorrectly) in the sciences. And Andrew is funny, with a good blend of story and instruction on what the hell statistics is about and how the most common techniques are applied. Even though it’s very basic, for instance he never goes into how to calculate a p-value, the conceptual and intuitive explanations for simple techniques was incredibly useful. Even the very beginning, looking at the differences between mean/standard deviations and median/quartiles was interesting.

Major things I remember: it doesn’t make sense to calculate multiple p-values for a single experiment and definitely never compare them. p-values is an indication of the likelihood that your null hypothesis is true, so calculating more than one implies you’re testing more than one null hypothesis. In addition, you just can’t compare them. I doesn’t make any sense. Also, p-values can be skewed by sample size, so make sure you understand your data before calculating a p-value otherwise it might give you a deceptive result.

He got a little more technical towards the end, but not as much as I’d like. He started talking about Wilcoxon values and ANOVA but didn’t really go into what they were. I think it would have been better if he got more technical as the book went on, where at the end the level was maybe college math. That way people could bow out of the book when the math got too involved, or continue reading but skip the mathematical explanations.

Still, a fun and easy read and probably one everyone should be reading in high school to ensure everyone is statistically literate.

Crafting the Personal Essay

By Dinty W. Moore.

Another book about writing personal essays. I did not like this book at first, or rather I did not feel like Dinty Moore was an appropriate expert to give me advice. This probably had something to do with his style or tone which simply did not jive with me, unlike Philip Lopate’s which, though I thought a lot of what he had to say was a bit silly, at least had more ethos to me. However I stuck it out and came to respect Dinty. Throughout the book he applies the lessons he is trying to teach to a sample essay of his. This felt a little contrived but the final essay, which he reveals at the end, is actually very good.

This book contains a lot of prompts, which I mostly ignored, and a couple sample essays, which I read.

His advice seems solid. He talked less about the expository nature than Philip did, more about retaining your reader’s interest and different styles of essays (lyrical, spiritual, humorous, etc.) However, he too thinks that most personal essays require the writer to be dealing with his thoughts and conflicts; he refers to this as ‘chasing mental rabbits’. In many ways the book is probably more useful to help people who aren’t sure what to write about, though he also gives a couple pithy case studies in students who have topics but refuse to include conflict or some other driving force.

It was a quick read and worth it, even if just to get a sense of what one of the best-selling books of this nature has to say. I’m not sure I’ll be reading more of these, though. Instead I will read some classic personal essays and try to tease them apart for guidance.