Shiny application for word correlation exploration.
Note wordcor will save smoothed version of the data in the current directory you run wordcor(). If your it in the same directory again it will reuse the smoothed versions and not recompute them.
The application will prompt you which data set to use. Currently there is six options of variations from the Google 1grams data set. Either the complete data set or restricted to the years after 1800. For both there are three options, original data set, relative scaling (each years relative frequency) and weighted relative frequency with additional scaling by the number of different words used per year.
The application user interface consists of multiple linked displays.
Double clicking will select the top 20 words with the most positive and most negative derivatives in that year.
The wavelet plot shows the sum of the wvaelet coefficients for all word (weighted by phase).
Selecting a region will select 40 words with the largest wavelet coefficients for that year and period length.
The primary and secondary selected word and the correlation to all other words represented on a line from -1 to +1 correlation.
The primary and secondary word can be changed through text input or clicking (for primary) or double clicking (for secondary) on a word represented on th line representation.
A correlation plot of with respect to primary and secondary.
Absolute Value selects whether positive and negative correlation should be distinguished or not.
Table shows selected words in the correlation plot, largest derivatives for the selected year, and largest wavelet cofficients for a selected region in the wavelet image.
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.