Social Insights — Sentiment & Daily Analytics
Daily social insights from NLP: we ingest Facebook, Instagram, and YouTube every 30 minutes, score multilingual sentiment, and deliver a daily report with alerts.
The Problem
Respect platform rate limits, handle noisy multilingual text (including emojis), and produce timely daily KPIs with automatic alerts when negativity spikes.
Our Solution
A Laravel dashboard with scheduled collectors every 30 minutes pulls platform data, normalizes it through DTO layers, and stores it in MySQL. A Python (Flask) API returns a sentiment label and score in multiple languages; scores are written back to posts and comments. Administration dashboards summarize trends, a daily email report lands in inboxes, and an alert fires when negative comments surge.
The Impact
Reliable 30-minute ingestion and normalization across Facebook, Instagram, and YouTube
Multilingual sentiment scoring (label + score) applied to posts and comments
Daily email report for stakeholders with top/bottom content
Real-time alerts when negative sentiment spikes
Consistent KPIs and trend windows for decision-making
Key Features
Top/bottom posts and trending topics
Segmentation by platform, timeframe, and keywords
Daily email reports with excerpts and links
Alerting on negative sentiment surges
Basic moderation tools and CSV exports
Challenges
Rate limits and API variability across platforms
Language and emoji normalization for robust scoring
On-time daily delivery with consistent KPIs
Lessons Learned
Cache recent windows to reduce API pressure and failures
Normalize and sanitize text before scoring
Precompute aggregates to keep reports fast and reliable
Co-design alert thresholds with business owners
Our Process
Assessment & Strategy
2 weeksDesign & Planning
2 weeksImplementation
8 weeksDeployment & Training
2 weeksProject Gallery
Want to be the next success story?
Let's discuss your project and turn it into an amazing digital reality