DEV Community

agenthustler
agenthustler

Posted on

How to Scrape YouTube Video Data with Python in 2026

#python #youtube #data #tutorial

YouTube is the world's second-largest search engine with over 2 billion monthly users. The platform generates more content data in a single day than most companies produce in a year — video titles, descriptions, view counts, engagement rates, comment sentiment, upload patterns, channel growth metrics.

For content strategists, marketers, and researchers, this data is gold. But getting it at scale? That's where things get complicated.

Let's look at what YouTube data actually enables — and the most practical way to extract it.

5 Use Cases for YouTube Data at Scale

1. Competitor Content Analysis

Your competitors are publishing on YouTube. Do you know what's working for them?

With structured channel data, you can:

  • See which video topics drive the most views in your niche
  • Track upload frequency and correlate it with subscriber growth
  • Identify which thumbnails and titles get the highest click-through (by comparing view counts relative to subscriber base)
  • Spot content gaps — topics your competitors haven't covered yet

A SaaS company analyzing their top 10 competitors' YouTube channels can build a data-driven content calendar in hours instead of weeks.

2. Trend Detection for Content Calendars

Trending topics on YouTube often predict broader cultural and market trends. By monitoring search results and video performance across categories, you can:

  • Detect rising topics before they hit peak saturation
  • Track view velocity (views per hour after upload) to identify viral content early
  • Map seasonal content patterns — what topics spike in Q1 vs Q4?
  • Build a "trending radar" that feeds into your editorial planning

3. Influencer Vetting for Brand Partnerships

Before you sign a $50K influencer deal, you need to verify the numbers. Follower counts can be inflated. Engagement can be bought. YouTube data lets you:

  • Pull actual view-to-subscriber ratios (healthy channels: 2-10%)
  • Check comment authenticity — are comments generic spam or genuine engagement?
  • Track performance consistency — are views steady or driven by one viral hit?
  • Compare engagement rates across multiple candidates objectively

4. Market Research Through Comment Sentiment

YouTube comments are an underused goldmine for market research. People share genuine opinions, complaints, and feature requests in video comments.

Extracting comments at scale lets you:

  • Run sentiment analysis across thousands of product review videos
  • Identify common pain points customers mention about your (or competitors') products
  • Track how brand perception shifts over time
  • Mine feature requests and use cases you hadn't considered

5. Academic and Media Research

Researchers studying misinformation, political content, or cultural trends need large YouTube datasets. Manual collection doesn't scale when you need data from thousands of channels.

Bulk extraction enables:

  • Cross-channel analysis of content patterns
  • Longitudinal studies tracking channel evolution over months or years
  • Large-scale content categorization and topic modeling

Why the YouTube API Falls Short

The official YouTube Data API v3 seems like the obvious choice. But in practice, it has hard limits that block serious data work:

Limitation Impact
10,000 quota units/day A single search costs 100 units — you get 100 searches per day
No bulk export Must request videos one at a time for detailed stats
OAuth required for some data Comment threads need authenticated requests
No historical trending data Can't look back at what was trending last month
Comment API restrictions Heavily rate-limited, no bulk access

If you need to analyze 50 competitor channels with 500+ videos each, you'll burn through your daily quota in minutes. Requesting a higher quota from Google requires a formal application and often takes weeks with no guarantee of approval.

Building a DIY scraper? YouTube's anti-bot protection is among the most aggressive on the web. Expect CAPTCHAs, IP blocks, and constant DOM changes.

The Practical Approach: Managed Extraction

Instead of fighting API quotas or building scrapers, use a managed solution that handles the complexity:

from apify_client import ApifyClient

client = ApifyClient("YOUR_APIFY_TOKEN")

# Pull top videos from a competitor channel
run_input = {
    "searchKeywords": ["saas marketing tips"],
    "maxResults": 100,
    "sortBy": "viewCount",
}

run = client.actor("cryptosignals/youtube-scraper").call(run_input=run_input)

for video in client.dataset(run["defaultDatasetId"]).iterate_items():
    title = video.get("title", "")
    views = video.get("viewCount", 0)
    likes = video.get("likeCount", 0)
    print(f"{title[:50]:50s} | {views:>10,} views | {likes:>6,} likes")
Enter fullscreen mode Exit fullscreen mode

What you get back — structured, analysis-ready data:

{
  "title": "How We Grew to $10M ARR with Content Marketing",
  "channelName": "SaaS Growth Hacks",
  "viewCount": 284500,
  "likeCount": 8420,
  "commentCount": 342,
  "publishedAt": "2026-01-15T14:00:00Z",
  "duration": "PT14M22S",
  "description": "...",
  "tags": ["saas", "content marketing", "growth"],
  "thumbnailUrl": "https://i.ytimg.com/vi/..."
}
Enter fullscreen mode Exit fullscreen mode

No API quota limits. No OAuth setup. No scraper maintenance.

Extracting Comments for Sentiment Analysis

Need the comment data too? Pair it with the comments scraper:

# Pull comments from specific videos
run_input = {
    "videoUrls": ["https://www.youtube.com/watch?v=VIDEO_ID"],
    "maxComments": 500,
}

run = client.actor("cryptosignals/youtube-comments-scraper").call(run_input=run_input)

for comment in client.dataset(run["defaultDatasetId"]).iterate_items():
    print(f"{comment.get('author', 'Anon')}: {comment.get('text', '')[:80]}")
Enter fullscreen mode Exit fullscreen mode

Real Example: Building a Competitor Content Dashboard

Here's how a content team might use this in practice:

from apify_client import ApifyClient

client = ApifyClient("YOUR_APIFY_TOKEN")

COMPETITOR_CHANNELS = [
    "competitor-channel-1",
    "competitor-channel-2",
    "competitor-channel-3",
]

all_videos = []
for channel in COMPETITOR_CHANNELS:
    run = client.actor("cryptosignals/youtube-scraper").call(
        run_input={"channelUrls": [f"https://youtube.com/@{channel}"], "maxResults": 50}
    )
    for video in client.dataset(run["defaultDatasetId"]).iterate_items():
        video["_channel"] = channel
        all_videos.append(video)

# Find top-performing content across all competitors
all_videos.sort(key=lambda v: v.get("viewCount", 0), reverse=True)

print("Top 10 Videos Across Competitors:")
for v in all_videos[:10]:
    print(f"  [{v['_channel']}] {v['title'][:45]} — {v['viewCount']:,} views")
Enter fullscreen mode Exit fullscreen mode

The scraper handles extraction. Your code handles the analysis. That's the right split.

Cost Comparison

Method Cost for 10K videos Quota/Rate Limits Maintenance
YouTube API v3 Free (but 100 searches/day cap) Severe Medium
DIY scraper $30-100/mo (proxies) Breaks frequently Very high
Apify actor ~$5-15 None None

Getting Started

  1. Sign up on Apify — free tier available
  2. Install the client: pip install apify-client
  3. Get your API token from Settings → Integrations
  4. Run the YouTube Scraper with your search terms or channel URLs
  5. For comments, use the YouTube Comments Scraper

You focus on what the data means for your strategy. Let the scraper handle getting it.

Using YouTube data for research or content strategy? Share your use case in the comments — I'm curious what people are building.

Top comments (0)