Social Media Analytics Dashboard

Project Overview

Built a comprehensive social media analytics platform that aggregates data from Twitter, Instagram, LinkedIn, and Facebook, providing real-time insights for marketing teams. This tool helps agencies prove ROI to clients by visualizing engagement metrics and sentiment trends across all channels in a single dashboard.

The Challenge

Marketing agencies often struggle with:

• Fragmented Data: Manually collecting stats from 4+ different platforms.
• Delayed Reporting: Client reports taking days to compile at the end of the month.
• Unclear Sentiment: Difficulty in gauging public reaction to campaigns instantly.
• Scalability: Handling millions of posts and comments during viral events.

Technical Solution

Architecture

The solution is a microservices-based application designed for high throughput and real-time analysis:

• Data Ingestion: Celery workers enabling asynchronous fetching from social APIs.
• Processing Engine: NLP service using Hugging Face transformers for sentiment analysis.
• Storage: Time-series data stored in PostgreSQL (TimescaleDB) for efficient querying.
• Frontend: A responsive React dashboard using D3.js for complex data visualization.

Key Features

1. Unified API Integration

   • Seamless connection to Twitter V2, Graph API (FB/Insta), and LinkedIn API.
   • Automatic token refresh and rate limit handling.

2. Real-time Sentiment AI

   • Classifies mentions as Positive, Neutral, or Negative with 90% accuracy.
   • Detects crisis situations (spikes in negative sentiment) and triggers alerts.

3. Automated Reporting

   • One-click PDF/PPT export for client presentations.
   • Scheduled email reports delivered every Monday morning.

4. Influencer Identification

   • Network analysis to identify key opinion leaders driving the conversation.

Technologies Used

• Backend: Python, FastAPI, Celery, Redis
• Frontend: React, TypeScript, Tailwind CSS, Recharts
• AI/ML: PyTorch, Hugging Face Transformers
• Database: PostgreSQL, Elasticsearch (for full-text search)
• DevOps: Docker, Kubernetes, GitHub Actions

Technical Highlights

Asynchronous Data Fetching

# Efficiently handling rate limits while fetching data
@celery_app.task(bind=True, max_retries=3)
def fetch_platform_data(self, platform_id, keywords):
    try:
        api = get_platform_client(platform_id)
        data = api.search(keywords, limit=1000)

        # Process and store asynchronously
        process_sentiment.delay(data)

    except RateLimitError as exc:
        # Smart backoff strategy
        wait_time = exc.reset_time - time.time()
        raise self.retry(exc=exc, countdown=wait_time + 5)

Results

Client Testimonial

“The automated insights from this dashboard changed how we pitch to clients. We can now show real-time impact of our campaigns instead of waiting for end-of-month spreadsheets.”

— Sarah Jenkins, Director at MarketingPro Agency