Trend Analysis

Trend analysis studies competitor YouTube Shorts to understand what works in your niche. It discovers patterns in hooks, tone, style, and pacing, then generates optimized configurations for your own channels.

What It Does

The trend analysis system tracks other creators in your niche, ingests their recent Shorts, runs AI analysis on each video to classify its style and strategy, tracks performance over time with metric snapshots, and synthesizes all of this data into actionable configuration that you can apply directly to your channels.

How It Works

The Full Cycle

Discover Channels --> Track --> Ingest Videos --> AI Analysis --> Snapshots --> Config Generation --> Apply to Channel

Step 1: Discover Channels

Search YouTube for channels in your niche. The system uses the YouTube Search API to find channels matching your query, then fetches their subscriber counts and upload playlist IDs.

Cost: 100 YouTube API quota units per search (daily limit: 10,000 units).

Step 2: Track Channels

Add discovered channels to your tracking list. Each channel is assigned a niche label (e.g., “gaming”, “finance”). The system stores their uploads playlist ID for later ingestion.

Step 3: Ingest Videos

Pull recent Shorts from tracked channels. The system:

• Fetches up to 50 recent uploads per channel
• Filters to Shorts only (duration 180 seconds or less)
• Skips videos already in the database
• Fetches auto-generated captions via YouTube’s timedtext API (no quota cost)
• Stores everything in the tracked_videos table

Step 4: AI Analysis

Each ingested video is analyzed by Claude (Anthropic’s AI) to classify:

• Hook type: question, shock stat, bold claim, nostalgia trigger
• Emotional tone: educational, shocking, funny, inspirational
• Visual style: talking head, stock footage, text overlay
• Music type: trending sound, background music, none
• Title pattern: number list, question, bold statement
• CTA type and placement: subscribe, comment, like — at start, end, or throughout
• Topic tags: 3-5 keyword tags

Results are stored in the video_analysis table.

Step 5: Performance Snapshots

The system takes metric snapshots at key moments after each video is published:

From consecutive snapshots, the system calculates view velocity (views per hour) and classifies each video’s trend:

Step 6: Config Generation

The system aggregates patterns from top-performing videos and synthesizes optimal parameters for your content:

• Best hook types and examples
• Optimal duration range and speaking pace
• Visual style and music recommendations
• CTA strategy
• Top-performing topics

This config is stored in the niche_configs table and can be automatically consumed by the script generator.

Step 7: Apply to Channel

The “Apply to Channel” button on the niche config page maps analysis results to your channel’s actual settings:

• visual_style: "stock_footage" enables Pexels slideshow
• music_type: "background_music" enables music provider
• text_density maps to subtitle style
• avg_duration maps to pipeline timing defaults

Where to Find It

• Trends Overview (sidebar) — Summary dashboard with getting-started guide
• Tracked Channels (sidebar) — Discover and manage channels to track
• Video Explorer (sidebar) — Browse tracked videos with analysis and velocity badges
• Niche Config (accessible from Trends Overview, per niche) — View and apply generated configs

Configuration

Trend analysis does not have per-channel configuration. It operates at the niche level, tracking channels across all niches you define. The generated niche configs can then be applied to any channel.

Dependencies

• YOUTUBE_DATA_API_KEY — For channel discovery and video ingestion (uses YouTube API quota)
• ANTHROPIC_API_KEY — For AI video analysis using Claude