YouTube has grown into a vast platform where users follow various channels, each representing different topics and interests. Managing these subscriptions and understanding what content you are mostly engaged with can be overwhelming. In this article, we'll walk through the creation of a Python application that analyzes a user's subscriptions and visualizes the distribution of topics.

Step 1: Authenticate with the YouTube API

First, we need to authenticate with the YouTube Data API v3 to access information about a user's subscriptions.

Install the Google Client Library

pip install --upgrade google-auth-oauthlib google-auth-httplib2 google-api-python-client

Authenticate and Save Credentials

import os
import pickle
import google_auth_oauthlib.flow
import googleapiclient.discovery

scopes = ["https://www.googleapis.com/auth/youtube.force-ssl"]

def get_authenticated_service():
    credentials_path = "credentials.pickle"
    if os.path.exists(credentials_path):
        with open(credentials_path, "rb") as f:
            credentials = pickle.load(f)
    else:
        flow = google_auth_oauthlib.flow.InstalledAppFlow.from_client_secrets_file(
            "client_secret.json", scopes)
        credentials = flow.run_console()
        with open(credentials_path, "wb") as f:
            pickle.dump(credentials, f)

    youtube = googleapiclient.discovery.build("youtube", "v3", credentials=credentials)
    return youtube

Make sure to download the client_secret.json file from your Google Cloud Console.

Step 2: Retrieve Subscriptions and Analyze Topics

After authentication, we can retrieve subscriptions and count the topics using a defaultdict.

Getting Subscriptions and Topics

from collections import defaultdict

def get_subscriptions(youtube):
    subscriptions = []
    next_page_token = None
    topic_count = defaultdict(int)

    while True:
        response = youtube.subscriptions().list(
            part="snippet",
            mine=True,
            maxResults=50,
            pageToken=next_page_token
        ).execute()

        subscriptions += response["items"]
        next_page_token = response.get("nextPageToken")

        if next_page_token is None:
            break

    for subscription in subscriptions:
        snippet = subscription["snippet"]
        channelId = snippet["resourceId"]["channelId"]
        response = youtube.channels().list(
            part="topicDetails",
            id=channelId
        ).execute()

        items = response.get("items", [])
        if items:
            topicDetails = items[0].get("topicDetails", {})
            topicCategories = topicDetails.get("topicCategories", [])
        else:
            topicCategories = []

        for topic in topicCategories:
            topic_count[topic] += 1

    return topic_count

Step 3: Visualize the Data with Matplotlib

Finally, we visualize the counted topics using Matplotlib.

Plotting the Topics

import matplotlib.pyplot as plt

def plot_topics(topic_count):
    topics, counts = zip(*topic_count.items())
    plt.bar(topics, counts)
    plt.xlabel('Topics')
    plt.ylabel('Counts')
    plt.title('Distribution of Topics in YouTube Subscriptions')
    plt.xticks(rotation='vertical')
    plt.show()

Running the Full Code

if __name__ == "__main__":
    youtube = get_authenticated_service()
    topic_count = get_subscriptions(youtube)
    plot_topics(topic_count)

Conclusion

We've created an application that authenticates with the YouTube API, analyzes subscriptions, and visualizes the distribution of topics. This is an excellent starting point for anyone interested in YouTube data analysis.

You can find the full code in the GitHub repository

kwisser / youtube-subscription-analyzer