YouTube Own Channel Analyzer

Analyze your YouTube channel's performance using the YouTube Data API v3.

Usage

/youtube-own-channel-analyzer @MyChannel
/youtube-own-channel-analyzer https://youtube.com/@MyChannel
/youtube-own-channel-analyzer UCxxxxxxxxxxxxxxxxxxxxxx --max-videos 300

Instructions

Step 1: Parse Arguments

• Channel (required): @handle, channel URL, or channel ID (UC...)
• --max-videos N (optional): how many recent uploads to analyze (default: 200, 0 = all)

Step 2: Get the API Key

Get a key from Google Cloud Console with YouTube Data API v3 enabled. Check Claude memory first; if not found, ask the user to paste it.

Step 3: Run the Bundled Script

Run scripts/analyze_channel.py — resolve the path relative to this skill's own directory:

YT_API_KEY=API_KEY python3 <skill-dir>/scripts/analyze_channel.py "@CHANNEL" [--max-videos N]

Dependency: pip3 install google-api-python-client.

Step 4: Read the Data

reports/data/channel-analysis-<channel-slug>-<YYYY-MM-DD>.json

Step 5: Write the Report

Write to the path the script printed: reports/channel-analysis-<channel-slug>-<YYYY-MM-DD>.md

What the Script Computes

Channel Resolution

@handle → channels.list(forHandle=...); /channel/UC... and bare UC... → channels.list(id=...); legacy /user/name → forUsername. Never search.list.

Video Collection

contentDetails.relatedPlaylists.uploads → playlistItems.list pagination → videos.list in batches of 50.

Content Types

Categorized by title/description/tag patterns:

Duration Buckets

Short (<5 min), Medium (5-15 min), Long (15-30 min), Very Long (30+ min).

Performance Metrics vs Benchmarks

The benchmarks ship inside the JSON as benchmarks so the report can compare directly.

Upload Patterns

Day-of-week and hour-of-day distribution, average days between uploads, uploads per week, and a consistency score (population stddev of the gaps — lower is more consistent).

Title Analysis

Rates of numbers, emojis, questions, brackets, ALL-CAPS words, and "how to"; average title length in characters and words; the 20 most common non-stopword terms.

Output Report Structure

# Channel Analysis Report: [Channel Name]
*Analyzed [date] | [N] videos*

## Executive Summary
- Subscribers, total views, videos, avg engagement, upload consistency
- One-line health verdict

## Channel Overview
- Basic info, statistics table, description, keywords, country, created date

## Content Analysis
- Category breakdown table (count, %, avg views, avg engagement)
- Duration distribution table
- Title patterns

## Performance Metrics
| Metric | This Channel | Benchmark | Verdict |
|--------|-------------|-----------|---------|
Compare each metric in `performance` against `benchmarks`.

## Upload Patterns
- Optimal day/hour, uploads per week, consistency stddev
- Day and hour distribution

## Engagement Analysis
- Top 5 high-engagement videos (`top_by_engagement`)
- Bottom 5 needing improvement (`needs_improvement`)
- Viral and underperforming counts

## Recommendations
- Content optimization, upload frequency, title suggestions
- Each recommendation tied to a specific number from the data

## Quota Usage
| Operation | Calls | Units |
|-----------|-------|-------|
Use the `quota_used.breakdown` block from the JSON.

Step 6: Report Completion

Tell the user the report path, subscriber/view/video counts, the health verdict, the top recommendation, and the quota consumed.

Quota Estimate

No search.list at all — this skill only touches its own channel.

Roughly 1 unit per 25 videos analyzed. The previously documented search.list(channelId=...) route cost 100 units per 50 videos (~400 units for the same 200 videos). Daily default allowance is 10,000 units, resetting at midnight Pacific Time.