Donald Trump Chat Bot

Goal

The goal of this project is to create a chat bot that will respond to the user in a manner that matches Donald Trump.

We are using python to collect and parse data collected from the internet.

Method

1. Use Selenium to scrape the internet for Trump's tweets and interview responses. This data is used to fill TrumpBot's vocabulary.

2. After finding sources, we used BeautifulSoup to find the desired html elements from the webpages and collected them into a text file.

3. We then passed the collected data to a python script that uses NLTK to separate each word and label it's part of sentence (i.e. noun, verb, adjective, etc.).

4. The data is then passed to another python script where it is cleared of any extraneous data such as urls or special characters. It is then organized by it's part of sentence (PoS) and iterations are added together.

Example of Adjectives: JJ:[('great', 118), ('many', 37), ('big', 32), ('much', 22), ('last', 21), ('good', 20), ('first', 19), ('american', 17), ('wonderful', 15), ('military', 15)]

1. A relation dictionary is made from a PoS bigram count. A list of 20 or less is created to be used as a random word list. A word will be randomly chosen based on the PoS and frequency of Trump's usage to create a sentence.

2. A user can now communicate with TrumpBot.

Data

HTML

All HTML files are stored in the html folder. The webpages were scraped to gather information for the chatbot.

HTML source files

• politico-trump.html
• twitter-trump.html

PYTHON

Python is used extensively to gather data from websites. Python scripts are then used to parse and clean the data from it's original source.

Libraries

• Selenium
• BeautifulSoup
• NLTK

TEXT FILES

The text files contain the parsed data from each step of data collection process.

Built With

• css
• html
• python