DecodeX
Home/Problem Statements/SIH26152
All 226 Statements
SIH26152SoftwareBlockchain & Cybersecurity
₹1,50,000 Prize (₹75k Joint)
•
Deadline: 20 September 2026
•
Submitted Ideas: 0/500

Social Media Analytics

Ministry / Organization: National Technical Research Organisation (NTRO)
SIH Official Portal
Technical Complexity
8/10
Dataset Accessibility
7/10
36h Hackathon Viability
9/10
Step 0 • Zero-Jargon Layman Summary

“This project involves building a digital 'listening post' that monitors social media to understand public mood, identify who is talking, and track how information spreads. It helps authorities detect emerging trends or potential unrest by analyzing conversations across platforms like X, Telegram, and Reddit in real-time.”

The DecodeX PS Decoder™

Deconstructed Problem Blueprint

Step 1

WHAT TO BUILD?

A centralized AI-driven intelligence dashboard that ingests multi-platform social media data to perform sentiment analysis, demographic profiling, trend detection, and network influence mapping.

Step 2

WHO IS IT FOR?

The National Technical Research Organisation (NTRO) and national security analysts who need to monitor digital narratives and identify key influencers in real-time.

Step 3

WHY DOES IT EXIST?

Social media is a massive, noisy, and unorganized source of information. Current manual monitoring is too slow to catch viral misinformation, radicalization, or emerging public sentiment shifts before they escalate.

Step 4

INPUT → PROCESS → OUTPUT ARCHITECTURE PIPELINE

01 • Inputs & Telemetry

Live API streams from X, Telegram, Reddit, and YouTube; historical post/comment datasets; user interaction metadata.

02 • Processing & AI Core

NLP for sentiment/sarcasm detection, demographic inference via profile/behavioral heuristics, graph theory for network topology, and time-series analysis for trend forecasting.

03 • Actionable Deliverables

Interactive dashboard with heatmaps, influencer network graphs, sentiment fluctuation charts, and automated alerts for anomalous trend spikes.

Step 5

CONSTRAINTS

  • API rate limits and compliance with platform Terms of Service (ToS)
  • Handling multilingual data (Indian regional languages)
  • Data privacy and anonymization (GDPR/DPDP Act compliance)
  • Real-time processing latency requirements
Step 6

JURY METRICS

  • Inference accuracy for sentiment and demographic classification
  • System throughput (posts processed per second)
  • Graph visualization responsiveness with large node counts
  • Latency between data ingestion and alert generation
Step 7

CORE BOTTLENECK

The 'Cold Start' problem in network analysis and the extreme difficulty of accurately detecting sarcasm/nuance in Indian regional languages and slang-heavy social media text.

Why You Should Choose This PS (Pros)

  • High national security relevance ensures strong jury interest
  • Clear, modular deliverables (Sentiment, Demographics, Trends, Network)
  • Excellent opportunity to showcase advanced AI/LLM integration

Challenges & Risks to Watch (Cons)

  • Strict API access limitations for X and Telegram
  • High computational cost for real-time graph processing
  • Risk of 'hallucination' in AI-based demographic inference
Hackathon Blueprint

Recommended Tech Stack & Grand Finale Features

Frontend & Mobile
React.jsRecharts / D3.js for graphsTailwind CSS
Backend & APIs
FastAPIPythonCelery for task queuing
AI / ML & Models
HuggingFace Transformers (BERT/RoBERTa for sentiment)Llama-3/DeepSeek for summarizationNetworkX for graph analysis
Database & Cloud
PostgreSQL + pgvectorNeo4j for graph dataRedis for real-time caching

Killer Features to Impress the SIH Jury

Cross-platform 'Influence Score' that tracks how a narrative jumps from Telegram to X
Automated 'Sarcasm Detector' specifically tuned for Indian English/Hinglish
Predictive 'Trend Velocity' indicator that warns of viral potential before it peaks
Privacy-first 'Anonymized Demographic Heatmap' to show regional sentiment without PII
View Official SIH 2026 Portal Raw Text
▼
•Background Social media platforms are complex ecosystems driven by human emotion, diverse demographics, and interconnected networks. To truly understand an online community, it is required look beneath the surface. This requires understanding how followers feel (Sentiment Analysis), who those followers are (Demographics), what topics are captivating them (Trend Tracking), and how they influence one another (Link Analysis). Combining these four vectors using AI is the key to unlocking true audience intelligence.
•Description Participants will be challenged to design and build an AI-driven Social Media Analytics Framework that processes raw platform data to extract deep, actionable audience insights. The system must leverage advanced Artificial Intelligence and Machine Learning techniques to simultaneously infer follower sentiment, map audience demographics, identify top trending narratives, and perform link/network analysis to uncover how information and influence flow among followers.
•Expected Solution AI solution must address the following five core components:
A.Continuous Data Collection & Timeline Management: Design a multi-platform data ingestion pipeline capable of pulling live data, posts, user interactions, and comments. The architecture must support a structured, time-stamped historical database to map out the exact chronology of conversations. The pipeline platform requirements are categorized as follows:
•Essentials (Must-Have): X (formerly Twitter) & Telegram.
•Desirable (Good-to-Have): Instagram & Facebook.
•Appreciable Additions: Reddit or YouTube (for extracting text-based context from video comments).
B.Multi-Dimensional Sentiment Inference: Use Natural Language Processing (NLP) to detect nuanced emotions (e.g., sarcasm, anxiety, excitement, supportive, against etc.) within user posts and comment threads, mapping how these sentiments fluctuate along the established data timeline.
C.Automated Demographic Profiling: Develop models to infer aggregate, anonymized follower demographics (such as age brackets, geographic distribution, language, and professional interests) based on public profile indicators, bio text, behavioral patterns etc.
D.Real-Time Trend & Topic Detection: Automatically identify, rank, and predict rising trends, viral keywords, and shifting discussions as they emerge chronologically in the dataset.
E.Link Analysis & Network Topology: Map the relationships among followers. Identify 'nodes of high influence' (key opinion leaders) and visualize how a trend or sentiment spreads from one user segment to another over time.