An experiment in concurrent text analysis and typesetting.
The cool thing about writing Python in DrawBot is that you can simultaneously analyze your data and typeset the results. This repo contains a Python script data_summarize.py that loops through a collection of books in HathiTrust, analyzes the text of those books, and prints the output in PNG and PDF. Each text analysis broadside provides a snapshot of some quantitative measures of the book.
The Python script is run in DrawBot, an open-source integrated development environment for Python often used by graphic designers for creating fonts and kinetic typography. Extracted Features is a data format devised by HathiTrust Research Center in order to enable text analysis on post-1926 books still under copyright.
When run in DrawBot, data_summarize.py takes the Extracted Features file for each of the ten books listed by HathiTrust ID in EF_files.txt. It then pulls the metadata for each book, takes a sample of random tokens, performs TF-IDF analysis, and calculates the appropriate font size for author name and title based on the letters that have to fit into a prescribed width.
Places are noted in the script where the broadside's design, font, and colors can be customized to suit the genre or period of the books you're analyzing.
 |
 |
 |
 |
- DrawBot, an open source IDE that unfortunately only runs on Mac. For an excellent guidebook to DrawBot, see Roberto Arista's Python for Designers
- the following Python libraries, installed via the command line with
pip install nltk htrc-feature-reader numpy sklearn pandas - Fonts: Futura and Hoefler Text will already be installed for Mac users. But I also use Fira Code, which will need to be downloaded and installed. Or you can specify any other fonts on your system.
Download a .zip of this repository, unzip, and open the folder. Open the DrawBot application. File -> Open -> data_summarize.py. Click "Run". The script will loop through each of the ten HathiTrust IDs listed in EF_files.txt, and should take about 45 seconds to finish. PNG and PDF broadsides for each book will be saved to the /output directory.
To analyze and print broadsides for different books, you can replace the HathiTrust IDs in EF_files.txt with any of the 17.1 million books that are included in the Extracted Features dataset. (Not every catalog record on HathiTrust will be included in Extracted Features.) Just search for your books on HathiTrust and find the HTID at the end of that book's URL, i.e. the HTID for https://babel.hathitrust.org/cgi/pt?id=uc1.32106016749274 would be uc1.32106016749274