Installation¶
Quick Install¶
This toolkit requires Python >= 3.10.
The fastest way to get started:
pip3 install convokit
That’s it! You’re ready to use ConvoKit. Alternatively, visit our Github Page to install from source.
Optional Dependencies¶
For specific features, you may need additional packages. ConvoKit relies on NLTK and SpaCy to implement certain NLP functions. If you have not already previously used these packages, they require additional first time setup:
For NLTK, download the punkt tokenizer: import nltk; nltk.download('punkt') (in a python interactive session)
For SpaCy, download the default English model: python3 -m spacy download en_core_web_sm
Troubleshooting¶
If you encounter difficulties with installation, check out our Troubleshooting Guide or ask in our Discord community.
Basic Usage¶
from convokit import Corpus, download
# Load a dataset
corpus = Corpus(download('conversations-gone-awry-corpus'))
# print corpus summary stats
corpus.print_summary_stats()
# Example: extract politeness features
from convokit import PolitenessStrategies
ps = PolitenessStrategies(verbose=5000)
corpus = ps.transform(corpus)
Next Steps¶
Datasets - Explore available datasets
Features & APIs - Discover analysis features
API Documentation - Detailed API reference