poketext

I am using text from Pokédex entries from main Pokémon games to practice text processing, web scraping, and using spaCy for NLP/NLU phenomena.

Notebooks present:

• scrape_pokedex_text_simple.ipynb: scrapes text from Bulbapedia Pokémon by Pokémon and pulls text from Pokédex entries from main series games, writing each entry to a text file.
• scrape_pokedex_text_bolstered.ipynb: does the same as above, but it scans the entries to check whether they actually use the name of the Pokémon. If they don't, the program replaces a suitable referring expression with the Pokémon's name.

Why Pokémon?

The Pokémon universe has been growing for over 20 years, and with that comes a surprising number of characters. This means the potential for a good amount of unique data, esp. re: the Pokémon themselves. However, it can also be easily constrained by only collecting certain kinds of examples (vs., eg, pulling random examples from blogs, social media, or forums).

• The massive amount of information about Pokémon lends itself nicely to a chatbot, and in fact, there already are some Pokémon chatbots, so I'd have something to compare my work to.
• The popularity of the series means that some of the data has already been encoded in usable formats (eg, pokemon.json, mentioned below, or Bulbapedia, where information pages about the Pokémon are nicely regular.
• Also, it adds a touch of lightness to the practice of learning to manipulate language data with code, as I'm a fan of the games myself!

Gratitude to:

• pokemon.json for the excellent Pokémon dataset that I pulled names from.

Copyright Notice

Please note the materials this project is based on are copyrighted by the Pokémon Company and its affiliates. All data gathered by the editors of Bulbapedia.