Back to Projects

AI-Enabled Climate & BRSR Dashboard

Unified analytics linking SEBI BRSR disclosures with national & global climate indicators—powered by Retrieval-Augmented Generation (RAG) for search, extraction, and auto-generated visuals.

Automotive OEM Pilot AI-RAG Benchmarking Extensible
AI-Enabled Climate & BRSR Dashboard

The Problem

Climate indicators (GHG inventories, energy mix, adaptation readiness) and corporate sustainability disclosures (SEBI's BRSR) live in silos. Policymakers, researchers, and industry leaders lack a single pane of glass to correlate company-level KPIs with national/sectoral climate context—slowing compliance tracking, benchmarking, and decarbonization planning.

Objectives

  • Visualize climate indicators (GHG, energy mix, adaptation) with sector filters and timelines.
  • Integrate BRSR data to assess sustainability performance of Indian industries.
  • Enable global and sector-specific comparisons (India vs peers; agriculture/energy/industry).
  • Add an AI-RAG agent that answers natural-language queries and auto-builds charts.

Methodology

  • Data Integration: parsers for BRSR PDFs; API/CSV pulls for climate/energy datasets; cleaning & normalization (units, timelines, terms).
  • Dashboard Design: responsive UI, filters by sector, timeframe, country/company.
  • Validation: data reliability tests, stakeholder UX sessions, speed optimization for large queries.
Architecture

End-to-end pipeline from data ingestion to interactive visualization

Data Sources

UNFCCC GHG Inventory
ND-GAIN Climate Index
IMF Climate Dashboard
Our World in Data
SEBI BRSR filings (NSE portal)
India Energy Dashboard (NITI Aayog)

Features

Combined View

Align corporate KPIs (e.g., Scope-1/2/3, water) with national energy/climate context.

Benchmarking

Compare companies vs domestic peers, global rivals, and sector baselines.

Query in Plain English

Ask "Top 5 auto OEMs by water consumption in 2024?" and get a chart + cited facts.

RAG Agent

The RAG agent retrieves relevant BRSR/climate facts from the integrated store, then composes answers grounded in those facts—reducing hallucinations and dynamically generating visuals (Matplotlib/Plotly) when needed.

  • Intent detection
  • Retrieval
  • Compose
  • Chart render
AI-Powered

Grounded reasoning with dynamic visualization

Sector Pilot & Scalability

Automotive OEM Pilot

Chosen for energy intensity, complex supply chains, fleet-emissions regulation.

Scalable Next

Minerals, Textiles, Petrochemicals, Cement, etc. (architecture is sector-agnostic).

Results & Insights

Fleet Emissions (Road Transport)

Heavy-duty trucks are the largest contributors despite lower counts; inventory built from VAHAN registrations using COPERT Tier-3 emission factors tuned to Indian conditions.

BEV vs ICEV (Lifecycle)

BEVs have higher production emissions (battery) but lower lifetime emissions; advantage increases as India's grid decarbonizes (e.g., ~943 → ~660 gCO₂/kWh by ~2030).

Deliverables

  • Interactive climate & BRSR analytics dashboard (sector-wise + global comparisons).
  • AI-powered RAG query agent with auto-charting.
  • Prototype dashboards (e.g., Power BI) and demo chatbot.
  • Methods write-up & presentation.

Attribution

Work done at CSIR-IIP (Climate Change & Data Science), with guidance from Dr. Tuhin Suvra Khan and Dr. Sunil Pathak; authored by Prateek Saxena (BITS Pilani).

Play Game