Quantitative analysis of the statements of the candidates for the election of the 44th student's parliament of Bonn
We want to analyze the statements of the seven lists candidating for the election of the 44th student's parliament of Bonn. These statements were published in the election magazine and each span four pages.
First, we have to extract the statements from the published pdf file.
For this we need to resort to optical character recognition (OCR) since some statements are included as page-spanning images. For the OCR, we use the python wrapper of tesseract. For the extraction of pdf pages as images we use pdf2image.
The ocr script can be found in ocr.py.
usage: OCR recognition of pdf pages [-h] [--path PATH] [--num-pages NUM_PAGES] [--lang LANG] first_page
Transforms pdf pages to Pillow Images using pdf2image and then uses tesseract for OCR. The recognized text is printed to stdout.
positional arguments:
first_page Number of the first page to consider
optional arguments:
-h, --help show this help message and exit
--path PATH Path of the pdf file (default: Wahlzeitung_2022.pdf)
--num-pages NUM_PAGES
Number of pages to consider, starting from first_page (default: 4)
--lang LANG three letter language code of the text to recognize (default: deu)
It can be invoked as
$ python ocr.py 34 > lp.logThe resulting text needs some (manual) processing, mostly due to misrecognized two-column layouts. You can finde our processed texts in the data folder. We decided to remove the names of individual candidates and their subjects (which are included in some statements).
to be added
The author is himself a candidate for the Liste Poppelsdorf.