Scraping of online reviews of cannabis strains and performing exploratory data analysis, including data cleaning, data transformations, and unsupervised association rule learning to mine associations between strain flavours and their medicinal properties.
Uses a combination of Selenium and BeautifulSoup + requests to do the scraping portion, followed by data manipulation and visualization with pandas and Plotly, and association rule learning with apriori.
To reproduce:
-
Clone the repository or download as ZIP and unzip it somewhere.
-
Run
link_crawl.pyusing Python 3. You'll need the Selenium Python bindings as well as the latest version of Chromedriver. -
Then run
strain_scrape_standard.pyto produce a CSV file with all the strain data nicely organized into columns and labelled. -
Finally, use Jupyter notebook to run
cannabis_strains_EDA.ipynb, where you can run individual cells, the entire notebook, or save the output to HTML. The HTML version of the notebook is available on my site.