Llm Public Opinion Analytics
Stand up a conversational public-opinion monitor that scrapes hot-search feeds, clusters topics with an LLM, scores sentiment, and pushes alerts.
Overview
llm-public-opinion-analytics is an agent skill most often used in Build (also Grow, Operate) that assembles multi-platform hot-search scraping, LLM clustering and sentiment, and multi-channel alerts into one public-opini
Install
npx skills add https://github.com/aradotso/data-skills --skill llm-public-opinion-analyticsWhat is this skill?
- Real-time scraping across 15 mainstream platforms and 26 ranking lists
- LLM-powered topic clustering, sentiment analysis, and video content extraction
- Multi-channel alerts via WeChat, Enterprise WeChat, Telegram, and email
- Conversational hot-search queries, topic search, and keyboard shortcuts for crawler control
- Prerequisites: Python 3.8+ and MySQL for the monitoring backend
Adoption & trust: 1 installs on skills.sh; 1 GitHub stars; 0/3 security scanners passed (skills.sh audits); trending (+100% hot-view momentum).
What problem does it solve?
You need continuous visibility into what audiences are saying across many hot-list platforms, but manual checks and ad-hoc scripts cannot scale to real-time clustering, sentiment, and alerting.
Who is it for?
Indie builders shipping audience-facing products in Chinese social ecosystems who want agent-guided setup of scrapers, LLM analytics, and alert channels.
Skip if: Teams that only need a single-platform API pull without persistence, LLM analysis, or operational alerting—or environments blocked from outbound scraping and third-party LLM calls.
When should I use this skill?
Analyze public opinion trends, scrape Chinese hot-search rankings, configure sentiment and clustering, or build a real-time public opinion monitoring system.
What do I get? / Deliverables
You get a MySQL-backed monitoring stack with conversational queries, LLM topic and sentiment analytics, and push notifications when narratives heat up—ready to extend with your own platforms and alert rules.
- Configured multi-platform scraper and ranking ingestion
- LLM clustering and sentiment analysis workflow
- Multi-channel alert configuration for hot topics
Recommended Skills
Journey fit
Spans multiple journey phases - primary shelf plus alternate fits below.
Canonical shelf is build/integrations because the skill centers on wiring scrapers, databases, LLM analysis, and notification channels into one assistant—not a one-off report. Integrations matches the 15-platform scraper layer, MySQL persistence, and multi-channel push hooks described in the overview.
Where it fits
Scan hot-search clusters around a problem niche before committing to a product angle.
Connect platform scrapers, MySQL storage, and LLM analysis into one assistant the agent can extend.
Watch narrative spikes that could affect launch timing or messaging on social hot lists.
Track sentiment shifts on branded or category topics after campaigns ship.
Keep Telegram or WeChat alerts firing when configured topics cross heat thresholds.
How it compares
Use as a full monitoring integration skill rather than asking the agent for one-off sentiment summaries in chat.
Common Questions / FAQ
Who is llm-public-opinion-analytics for?
Solo and indie builders who monitor trending topics and sentiment across many Chinese social and search hot lists and want that pipeline orchestrated through an agent skill.
When should I use llm-public-opinion-analytics?
During Build when wiring scrapers and LLM analytics into your product, during Grow when tracking campaign or brand narratives, and during Operate when you need ongoing hot-topic alerts and clustered topic views.
Is llm-public-opinion-analytics safe to install?
It implies network scraping, database credentials, and notification API secrets—review the Security Audits panel on this Prism page and lock down keys, crawl policies, and data retention before production use.
SKILL.md
READMESKILL.md - Llm Public Opinion Analytics
# LLM-Based Public Opinion Analytics Assistant > Skill by [ara.so](https://ara.so) — Data Skills collection. ## Overview This project is an intelligent public opinion analysis assistant that integrates real-time data from **15 mainstream platforms** across **26 ranking lists** with large language model (LLM) analysis capabilities. It provides conversational hot search queries, topic-specific searches, topic clustering, and sentiment analysis. The system supports: - Real-time web scraping from platforms like Weibo, Bilibili, Douyin, Baidu, etc. - LLM-powered content analysis (including video content extraction) - Multi-channel push notifications (WeChat, Enterprise WeChat, Telegram, Email) - Keyboard shortcuts for crawler control - Quick data lookup and platform jumping ## Installation ### Prerequisites 1. **Python Environment**: Python 3.8+ 2. **MySQL Database**: MySQL 5.7+ or 8.0+ 3. **Browser Driver**: ChromeDriver or EdgeDriver ### Step 1: Browser Driver Setup Download the driver matching your browser version: - **Chrome**: [ChromeDriver Downloads](https://chromedriver.chromium.org/) - **Edge**: [EdgeDriver Downloads](https://developer.microsoft.com/en-us/microsoft-edge/tools/webdriver/) Add the driver to your system PATH: ```bash # macOS/Linux export PATH=$PATH:/path/to/driver/directory # Windows: Add to System Environment Variables ``` Verify installation: ```bash chromedriver --version # or msedgedriver --version ``` ### Step 2: Clone and Install Dependencies ```bash git clone https://github.com/hmmnxkl/LLM-Based-Intelligent-Public-Opinion-Analytics-Assistant.git cd LLM-Based-Intelligent-Public-Opinion-Analytics-Assistant # Create virtual environment python -m venv venv source venv/bin/activate # On Windows: venv\Scripts\activate # Install dependencies pip install -r requirements.txt ``` ### Step 3: Database Setup Create MySQL database and tables: ```python # Reference init.py for schema import mysql.connector conn = mysql.connector.connect( host=os.getenv('MYSQL_HOST', 'localhost'), user=os.getenv('MYSQL_USER'), password=os.getenv('MYSQL_PASSWORD') ) cursor = conn.cursor() cursor.execute("CREATE DATABASE IF NOT EXISTS hotsearch_db CHARACTER SET utf8mb4 COLLATE utf8mb4_unicode_ci") cursor.execute("USE hotsearch_db") # Create tables (see init.py for full schema) cursor.execute(""" CREATE TABLE IF NOT EXISTS hot_search_items ( id INT AUTO_INCREMENT PRIMARY KEY, platform VARCHAR(50), title VARCHAR(500), url TEXT, rank_index INT, heat_value VARCHAR(100), collected_at DATETIME, content TEXT, sentiment VARCHAR(20), INDEX idx_platform (platform), INDEX idx_collected (collected_at) ) """) conn.commit() ``` ### Step 4: Environment Configuration Create `.env` file in project root: ```bash # MySQL Configuration MYSQL_HOST=localhost MYSQL_PORT=3306 MYSQL_USER=your_mysql_user MYSQL_PASSWORD=your_mysql_password MYSQL_DATABASE=hotsearch_db # LLM Configuration (OpenAI-compatible API) OPENAI_API_KEY=your_api_key OPENAI_API_BASE=https://api.openai.com/v1 MODEL_NAME=gpt-4 # Or use Huawei Pangu Model (local deployment) # PANGU_MODEL_PATH=/path/to/pangu/model # PANGU_API_URL=http://localhost:8080 # Push Notification Channels # WeChat Work Bot WECHAT_WORK_BOT_WEBHOOK=your_webhook_url # WeChat Work App W