What do you think?
Rate this book
Graphical Methods for Data Analysis
This book present graphical methods for analysing data. Some methods are new and some are old, some require a computer and others only paper and pencil; but they are all powerful data analysis tools. In many situations, a set of data � even a large set- can be adequately analysed through graphical methods alone. In most other situations, a few well-chosen graphical displays can significantly enhance numerical statistical analyses.
336 pages, Hardcover
First published January 1, 1983
31 people want to read
Ratings & Reviews
Friends & Following
Create a free account to discover what your friends think of this book!
Community Reviews
Displaying 1 of 1 review
July 13, 2018
TODO full review:
+ Classic book on information visualization. Main strength: discusses many types of graphs developed for exploring the statistical properties of datasets encountered by the Bell Labs team in the 1970s and 1980s. Many of these types are no longer in use, but the design ideas behind them are still worth understanding.
+ Covers basic graph types and their statistical meaning, plus pitfalls in interpretation.
+ Some of the graph types now extinct: Q(.xy) graphs (Section 2.2), 1D scatterplots (Section 2.4), notched box plots (Section 3.4), sunflower plots (a form of pixelation, Section 4.9), and profile symbol plots and Kleiner-Hartigan trees (Section 5.4). Some of the unfortunately not yet extinct: star symbol plots (now known as spider charts or Kiviat diagrams, Section 5.4).
+ Chapter 8 covers general principles and ideas about generating meaningful plots about quantitative data. Among the principles: iteration (no data visualized in just one plot or even in just one go), interpretability (relatively easy when data has meaning directly related to physical reality, but especially important for derived and combined plots that represent subtle effects that are more difficult to relate to immediately identifiable reality), flexibility (of proposed graphing techniques, to accommodate a variety of situations), and true message (avoiding to delude ourselves, and avoid to delude others -- guess the Soviet Russian Stats Office employees have not read this book, back in the days).
+ Classic book on information visualization. Main strength: discusses many types of graphs developed for exploring the statistical properties of datasets encountered by the Bell Labs team in the 1970s and 1980s. Many of these types are no longer in use, but the design ideas behind them are still worth understanding.
+ Covers basic graph types and their statistical meaning, plus pitfalls in interpretation.
+ Some of the graph types now extinct: Q(.xy) graphs (Section 2.2), 1D scatterplots (Section 2.4), notched box plots (Section 3.4), sunflower plots (a form of pixelation, Section 4.9), and profile symbol plots and Kleiner-Hartigan trees (Section 5.4). Some of the unfortunately not yet extinct: star symbol plots (now known as spider charts or Kiviat diagrams, Section 5.4).
+ Chapter 8 covers general principles and ideas about generating meaningful plots about quantitative data. Among the principles: iteration (no data visualized in just one plot or even in just one go), interpretability (relatively easy when data has meaning directly related to physical reality, but especially important for derived and combined plots that represent subtle effects that are more difficult to relate to immediately identifiable reality), flexibility (of proposed graphing techniques, to accommodate a variety of situations), and true message (avoiding to delude ourselves, and avoid to delude others -- guess the Soviet Russian Stats Office employees have not read this book, back in the days).
Displaying 1 of 1 review