📋 5,347	🧠 40+	🏢 1,500+	🛡️ 0.6%	📊 7	💰 99.4%
Job descriptions	Canonical tech skills	Hiring companies	Actually disclosed salaries	Premium modules	Transparent salary modeling

---

📌 Business Questions Answered

• What skill patterns are associated with higher modeled compensation in this dataset?
• How much of the job market data online is actually factual versus blindly modeled?
• What is the precise tech-stack difference between Product-tier and Consulting-tier companies?
• Which technologies have the strongest co-occurrence frequency (e.g. AWS + Snowflake)?
• What is my exact learning roadmap to break the ₹16L compensation ceiling?

---

🔍 Key Findings (from real data)

• Cloud Multiplier (Model-Based): The salary estimation model assigns 1.15×–1.25× premiums for Cloud tools (Snowflake, Databricks, dbt), sourced from AmbitionBox/Glassdoor benchmarks. Note: 0 of 33 disclosed salaries in this dataset included cloud skills, so this is a model assumption.
• The "Dirty Data" Truth: Only 33 out of 5,347 (0.6%) job descriptions possessed explicitly disclosed salaries, showing that salary disclosure is extremely sparse in this dataset.
• Product Tier Premia (Model-Based): The model applies a 1.25× multiplier for Product-tier companies vs 1.10× for Consulting-tier (based on external benchmarks). 0 of 33 disclosed salaries came from either tier, so this is an assumption, not an observed finding.
• Top Tech Target: General programming capability (Python) combined with heavy data manipulation (SQL) retains absolute market dominance across 31% of total listings.

Data Honesty: Every metric displayed traces strictly to the custom Python NLP pipeline. Only 0.6% of salaries are observed (disclosed); all remaining salary figures are model-estimated using external benchmarks and clearly labelled. Salary model parameters (skill premiums, tier multipliers) are transparent assumptions, not disguised as observed data.

---

🛠️ Tech Stack

Layer	Technology
Dashboard Framework	Vite
Frontend Mechanics	Vanilla JavaScript (ES6)
Visualizations	Chart.js 4.x + Glassmorphism Design System
Data Processing	Python · Pandas · NumPy
Extraction Pipeline	Regex + Custom Entity-Matching Hash Tables
Data Transfer Payload	Pre-compiled multidimensional JSON blocks

The entire system is uncoupled. The heavy Python processing runs offline to calculate dashboard_data.json, allowing the Vite web application to render with zero latency.

---

📊 Dashboard Pages

🎛️ Command Center — Hero Telemetry & Data Trust Layer	💰 Salary Intelligence — Interactive Pay Simulator
🎯 Skill Demand Radar — Category-Segmented Prevalence	🏢 Company War Room — 1,500+ Employer Matrix
🔗 Skill Synergy Map — Co-occurrence Correlation	🗺️ Career Pathfinder — Auto-Generated ROI Roadmap
📰 Market Pulse Report — Data Freshness & 2019 vs 2024 Comparison

Page	What It Shows
🎛️ Command Center	Live market telemetry, top employer tracking, and our audited Trust badge
🎯 Skill Demand Radar	Visual prevalence mapping mapping DBs vs Analytics vs Programming toolkits
💰 Salary Intelligence	Simulator projecting lifetime LPA trajectory across 5 experience tiers
🏢 Company War Room	Target searching across 1,500+ active hiring entities instantly
🔗 Skill Synergy Map	Co-occurrence patterns showing which software combinations appear together most frequently
🗺️ Career Pathfinder	Checkbox assessment that calculates the single missing tool driving the most ROI
📰 Market Pulse	Automated, export-ready executive briefings

---

🔢 Key Numbers — Pipeline-Verified

 5,347   total scraped Analyst & Data Scientist JD records
    41   canonical target tools monitored (SQL, Tableau, dbt, etc.)
 ~200+   semantic synonyms collapsed via rigorous text processing
    33   jobs containing disclosed, factual compensation (0.6%)
 5,314   jobs completed via algorithmic proxy benchmarks (99.4%)
 5,116   rows with explicitly identifiable corporate entities

---

🚀 Run Locally

git clone https://github.com/Yashaswini-V21/TalentPulse-Engine.git
cd TalentPulse-Engine/src

1. Launch the Dashboard (Vite):

cd webapp
npm install
npm run dev
# Dashboard launches at http://localhost:5173

2. Rebuild the Intelligence Core (Python):

# Return to /src
pip install -r requirements.txt
python build_pipeline.py
python enrich_salary.py
python build_dashboard_json.py  # compiles output payload to /webapp

---

📁 Project Structure

TalentPulse_Engine/
│
├── 📁 src/
│   ├── 📄 build_pipeline.py           ← Primary NLP dataset parsing
│   ├── 📄 enrich_salary.py            ← Honest proxy-filling script (safeguards real data)
│   ├── 📄 build_dashboard_json.py     ← Package compiler pushing CSVs to JSON payload
│   ├── 📄 requirements.txt            ← Pinned Python dependencies
│   │
│   ├── 📁 data/
│   │   ├── raw/                       ← Origin Dataset CSVs
│   │   └── clean/                     ← Output analysis aggregations
│   │
│   ├── 📁 nlp/
│   │   └── skill_extractor.py         ← Vocabulary logic mapping and custom dictionaries
│   │
│   ├── 📁 sql/
│   │   └── practice_queries.sql       ← Embedded analytical logic equivalents
│   │
│   └── 📁 webapp/                     ← Ultra-light frontend application
│       ├── 📄 index.html              ← Entrypoint 
│       ├── 📄 main.js                 ← Handles 7 distinct visual modules 
│       ├── 📄 style.css               ← Native glassmorphism token variables
│       └── 📁 public/                 
│           └── dashboard_data.json    ← Pre-compiled multidimensional JSON from Python
│
├── 📁 assets/                         ← Premium dashboard screenshots
└── 📄 README.md                       ← Project overview

---

👤 Author

Yashaswini V · LinkedIn · GitHub

---

   

_{TalentPulse Engine  ·  MIT Licensed  ·  Enterprise Intelligence Built Correctly  ·  2026}