HankComments

Hank Green had a video, where he asked his viewers to recommend books, movies, TV shows, and video games, that they thought were calming. I iterated through all the comments, stored them in a txt file, then used GPT 3.5 Turbo to parse the comments into a list of recommendations, grouped by category. I then used GPT 4 Code Interpreter to analyze the recommendations, and create visualizations for each category.

Animated Visualization

Here's an animated overview of the top recommendations across all categories:

Detailed Static Graphs

For a more detailed look at each category, check out the static visualizations below:

📚 Books (Click to expand)

🎬 Movies (Click to expand)

📺 TV Shows (Click to expand)

🎮 Video Games (Click to expand)

How It Works

1. Collection: YouTube comments were collected from Hank Green's video asking for calming media recommendations.
2. Processing: Comments were processed using GPT-3.5 Turbo to extract and categorize recommendations.
3. Analysis: Data was cleaned and analyzed to identify the most frequently mentioned items in each category.
4. Visualization: Visualizations were created to showcase the findings.

Installation

1. Clone this repository:

   git clone https://github.com/yourusername/HankComments.git
   cd HankComments

2. Install the required dependencies:

   pip install -r requirements.txt

3. Create a .env file with your API keys (see .env.example for format):

   cp .env.example .env

   Edit the .env file with your API keys.

Usage

Run the complete pipeline with a single command: