🎯 AI Resume Analyzer & Job Match System

Nixon Paul [21MIC0159]

An NLP-powered resume analyzer and job matcher built with Python and Streamlit.
Upload a resume, paste a job description, and get instant skill matching, ATS scoring, and improvement suggestions.

---

✨ Features

Feature	Description
📄 Resume Parsing	Supports PDF, DOCX, and TXT formats
🔬 NLP Pipeline	Tokenization → Stopword Removal → Lemmatization
🛠️ Skill Extraction	Matches 120+ skills across 10 categories
📊 TF-IDF Matching	Cosine similarity for resume–JD comparison
🎯 ATS Score	Weighted score (similarity + skills + completeness)
💡 Suggestions	Actionable improvement recommendations
📈 Visualizations	Gauge charts, bar charts, donut charts (Plotly)

---

🏗️ Project Structure

NLP-based-Resume-Analyzer/
├── app.py                  ← Main Streamlit app
├── requirements.txt        ← Python dependencies
├── setup.sh                ← One-command setup script
│
├── data/
│   └── skills.csv          ← Skill dataset (120+ skills)
│
├── utils/
│   ├── parser.py           ← Resume text extraction (PDF/DOCX/TXT)
│   ├── preprocess.py       ← NLP preprocessing pipeline
│   ├── skill_extractor.py  ← Skill matching engine
│   ├── similarity.py       ← TF-IDF cosine similarity
│   ├── ats_score.py        ← ATS score computation
│   └── suggestions.py      ← Recommendation engine
│
└── sample_data/
    ├── resume_software_engineer.txt
    ├── resume_data_scientist.txt
    ├── jd_backend_engineer.txt
    └── jd_data_scientist.txt

---

⚙️ Setup & Installation

Step 1 — Clone / Open the project

cd "NLP-based-Resume-Analyzer"

Step 2 — Run setup (installs all dependencies)

bash setup.sh

This will:

1. Install all Python packages from requirements.txt
2. Download the spaCy English model (en_core_web_sm)
3. Download required NLTK data (stopwords, punkt, wordnet)

Step 3 — Launch the app

streamlit run app.py

The app opens at http://localhost:8501

---

🧪 Manual Testing Guide

Use the provided sample files in sample_data/ to test:

Test 1 — High Match (Expected ~75–90%)

• Resume: resume_software_engineer.txt
• JD: jd_backend_engineer.txt

Test 2 — High Match (Expected ~80–92%)

• Resume: resume_data_scientist.txt
• JD: jd_data_scientist.txt

Test 3 — Lower Match (Expected ~30–55%)

• Resume: resume_software_engineer.txt
• JD: jd_data_scientist.txt

---

🔬 NLP Pipeline

Raw Resume Text
    ↓ Lowercase conversion
    ↓ Punctuation & special char removal
    ↓ Tokenization (spaCy)
    ↓ Stopword removal (NLTK + spaCy)
    ↓ Lemmatization (spaCy)
    ↓ Skill phrase matching
    ↓ TF-IDF vectorization (scikit-learn)
    ↓ Cosine similarity computation
    ↓ Weighted ATS score
    ↓ Dashboard visualization

---

📐 ATS Score Formula

Component	Weight
TF-IDF Cosine Similarity	50%
Skill Overlap Ratio	30%
Resume Completeness Signals	20%

---

🛠️ Tech Stack

Layer	Technology
UI	Streamlit
NLP	spaCy, NLTK
ML	scikit-learn (TF-IDF)
Visualization	Plotly
File Parsing	pdfplumber, python-docx
Language	Python 3.9+

---

📋 Requirements

• Python 3.9 or higher
• Internet connection for initial setup (model downloads)
• ~500 MB disk space for NLP models