Graphics Reference
In-Depth Information
For example, global trade flows between countries can be represented as a
table of numbers (
http://stats.oecd.org
). Each cell represents a link, with
the first column and first row indicating nodes:
From/To, Austria, Belgium, Denmark, ...
Austria, n/a, 2.197b, 1.014b, ...
Belgium, 4.411b, n/a, 3.681b, ...
Denmark, 0.753b, 1.284b, n/a, ...
...
See the
Trade Flow
example in the Supplementary Materials on this
book's companion website. In the spreadsheet, you can see a technique
for transforming the matrix of links into a list of links using spreadsheet
formulas.
Statistical Correlation (for Example, Stocks, News Stories)
In all the previous examples, a connection existed in the data. Graphs can
also be created statistically.
For example, two stocks can be said to have a strong correlation if their
prices move up and down together. This can be computed statistically as a
function of the two time series of data. Raw price time series may look like
this:
Stock, Jan-2, Jan-3, Jan-4, Jan-5,...
AAPL, 520.21, 515.98, 518.22, 514.29
GOOG, 958.37, 968.77, 978.11, 988.33
IBM, 177.34, 176.33, 175.44, 176.58
...
Links can be formed between every pair of time series using a correlation
function (that is, stock prices that move up and down similarly are highly
correlated, whereas stocks that move in opposite directions are inversely
correlated), as shown here:
Stock1, Stock2, 1 year correlation
AAPL, GOOG, 0.94
AAPL, IBM, 0.77
GOOG, IBM, 0.66
...
