thanhtoan1196 / python-fanpage-scraping

A small project to crawl public posts from Facebook Fanpage without Facebook Graph API

A small project aims at crawling public facebook fanpage

Requirements

*python >= 3.6 (some dependencies may not work properly with newer versions, at the time I made this project, underthesea module is not loaded with the latest python version 3.8.3)
*Tesseract client (see: https://github.com/tesseract-ocr/tesseract)
*pytesseract:

pip install facebook-scraper

underthesea: (see https://github.com/undertheseanlp/underthesea)

wordcloud:

pip install wordcloud

Project structure

|-- .gitignore
|-- sample
    |-- _processed.json
    |-- chinh_tri_xa_hoi.csv
    |-- processed.json
    |-- stat-full.csv
    |-- stat.csv
|-- count_reactions_by_page.py
|-- csv.py
|-- datahandling.py
|-- facebook_scraper.py
|-- mask.png
|-- nlp_processing.py
|-- README.rst
|-- scraping.py
|-- vietnamese-stopwords.txt
|-- wordcloud.py
|-- wordtokens_stat.py

1. Installation

The facebook_scraper.py in this project is a modified version of Kevinzg facecbook-scraper (see at: https://github.com/kevinzg/facebook-scraper) You can either move this file to your current python's site_packages folder or install the original module and replace the content of this main module file by this one.

If there is any better workaround, go for it cuz I'm not a pythonworm :(((

For other modules, install all the packages listed in the Requirements section above. You better go to the link I provided to get them all installed correctly.

2. Data Scraping

Usage

• Open scaping.py and change the final line of code

...     extract_page_public_posts('<page name>', <quantity>)

3. Text cleaning and Image Text extraction

Usage

• Open datahandling.py and change the code in main method

...     try:
...     pool = []
        #Replace your data from scaping.py in the below code
        pool = pool + load_data('./dummies1.json', 'DM')
        pool = pool + load_data('./dummies2.json', 'DN')
        pool = pool + load_data(<path to fanpage data from scraping.py>, <tag name for fanpage data>)

4. Sentiment and Classification

Usage

• Run nlp_processing.py
• You don't need to modify any line of code provided that you did't change the output file location or name in step 3.

5. Export to csv

Usage

• Run export_to_csv.py
• You don't need to modify any line of code provided that you did't change the output file location or name in step 4.
• With the result of this file, you can use any sort of visualization applications or services having csv supported file to analyze your data.

6. Word Statistics

Usage

• Open wordtokens_stat.py and change the final line of code
• Default data file location is: processed.json
• To get topic names: open processed.json and copy the topic name wrapped inside ""

...   word_tokens_stat(<path to received file from step 4>, '<topic name>')

7. WordCloud

Usage

• Open wordcloud.py and change the final line of code
• Default data file location is: processed.json
• To get topic names: open processed.json and copy the topic name wrapped inside ""

...   word_cloud(<path to received file from step 4>, '<topic name>')