Political Video Comment Analysis Tool
Budget: ₹750 – ₹1,250 INR
We are building a lean MVP to research public opinion by analysing comments on political videos posted on social media. The tool should include comment scraping, AI-based opinion segmentation, and visual reporting. In a second phase, the tool will also process video transcripts for deeper AI context understanding.
To accelerate delivery and reduce cost, we encourage the use of no-code or low-code platforms for workflows, integrations, and dashboards where possible.
Objectives
Phase 1 – MVP
1. Scrape Comments
Target platforms: TikTok, Facebook, YouTube
Support input via:
Hashtag/topic
Manual list of video URLs
Use ready-made scraping tools (e.g. Apify, Octoparse) for fast delivery
2. Store Data
Save all comments and metadata in a structured database
(e.g. MongoDB, PostgreSQL, or Airtable)
Metadata includes:
Video ID
Comment text
Likes
Timestamp
Username (if available)
3. AI Analysis & Segmentation
AI agent groups comments into public opinion categories
Example output:
"45% support the message"
"32% are negative or critical"
"23% show confusion or questions"
Each group includes:
Segment name
Comment count
Percentage
Sample quotes
4. Report Output
Web-based dashboard or no-code UI
(e.g. Retool, Softr, Glide)
Visuals:
Pie or bar charts for sentiment breakdown
Tables with segmented comment summaries
Export options (CSV or PDF)
5. Scalability
Must support processing up to 500,000+ comments
Support batching, pagination, or background processing
6. Training Loop (Optional for Phase 1)
Manual review and correction of AI classifications
Improve AI performance with human feedback
Phase 2 – Transcript Integration
1. Scrape Video Transcripts
Extract captions or transcripts from platforms (e.g. YouTube)
Use platform APIs or third-party tools
2. AI Contextual Processing
Include transcripts in AI processing pipeline
Provide better context for analysing comments
Technical Considerations
Area
Suggested Tools / Notes
Scraping
Apify, Octoparse, or Puppeteer/Selenium
AI
OpenAI API (GPT-4), Claude, or other LLMs
Workflows
Make, Zapier, LangChain (if needed)
DB
Airtable (fast start), PostgreSQL, MongoDB
UI
Retool, Softr, Glide, or custom Streamlit app
Request for Estimate
Please include in your response:
Timeline for Phase 1 (and optionally Phase 2)
Budget estimate
Tool/platform recommendations
Any similar past work or references
To accelerate delivery and reduce cost, we encourage the use of no-code or low-code platforms for workflows, integrations, and dashboards where possible.
Objectives
Phase 1 – MVP
1. Scrape Comments
Target platforms: TikTok, Facebook, YouTube
Support input via:
Hashtag/topic
Manual list of video URLs
Use ready-made scraping tools (e.g. Apify, Octoparse) for fast delivery
2. Store Data
Save all comments and metadata in a structured database
(e.g. MongoDB, PostgreSQL, or Airtable)
Metadata includes:
Video ID
Comment text
Likes
Timestamp
Username (if available)
3. AI Analysis & Segmentation
AI agent groups comments into public opinion categories
Example output:
"45% support the message"
"32% are negative or critical"
"23% show confusion or questions"
Each group includes:
Segment name
Comment count
Percentage
Sample quotes
4. Report Output
Web-based dashboard or no-code UI
(e.g. Retool, Softr, Glide)
Visuals:
Pie or bar charts for sentiment breakdown
Tables with segmented comment summaries
Export options (CSV or PDF)
5. Scalability
Must support processing up to 500,000+ comments
Support batching, pagination, or background processing
6. Training Loop (Optional for Phase 1)
Manual review and correction of AI classifications
Improve AI performance with human feedback
Phase 2 – Transcript Integration
1. Scrape Video Transcripts
Extract captions or transcripts from platforms (e.g. YouTube)
Use platform APIs or third-party tools
2. AI Contextual Processing
Include transcripts in AI processing pipeline
Provide better context for analysing comments
Technical Considerations
Area
Suggested Tools / Notes
Scraping
Apify, Octoparse, or Puppeteer/Selenium
AI
OpenAI API (GPT-4), Claude, or other LLMs
Workflows
Make, Zapier, LangChain (if needed)
DB
Airtable (fast start), PostgreSQL, MongoDB
UI
Retool, Softr, Glide, or custom Streamlit app
Request for Estimate
Please include in your response:
Timeline for Phase 1 (and optionally Phase 2)
Budget estimate
Tool/platform recommendations
Any similar past work or references