ResumeParser: Intelligent Resume Screening & Analysis Tool

📘 Abstract

ResumeParser is a web-based application that simplifies and automates the resume screening process using Natural Language Processing (NLP) and AI-powered analytics. Designed for recruiters, placement teams, and HR departments, the system extracts structured information from unstructured PDF resumes and delivers a detailed candidate profile with ATS (Applicant Tracking System) scoring, keyword matching, and skill distribution insights.

---

🏗️ System Architecture

The application is built using a modular layered architecture, split across:

🔹 1. Frontend (Client Layer)

• Technology: HTML, CSS, JavaScript, Chart.js, D3.js
• Main File: index.html
• Role:
  • Allows users to upload multiple PDF resumes
  • Provides dynamic UI for:
    • Resume preview using <iframe>
    • ATS score visualization (doughnut chart)
    • Word cloud using D3
    • Skill distribution chart
    • Modal popup for detailed analysis

🔹 2. Backend (Application Layer)

• Technology: Python, Flask
• Main File: app.py
• Role:
  • Handles file uploads and stores them in uploads/ directory
  • Extracts raw text from PDFs using pdfminer
  • Uses spaCy (en_core_web_sm) for NLP-based:
    • Name extraction (using proper nouns)
    • Email and phone number extraction (Regex)
    • Skill detection from predefined keywords
  • Computes ATS compatibility using weighted scoring
  • Generates detailed analysis with:
    • Education
    • Work Experience
    • Projects
    • Certifications
    • Soft and Technical Skills
  • Prepares data for:
    • Charts
    • Word cloud
    • Resume preview
    • Candidate summary

🔹 3. Data Layer (PDF Input & JSON Output)

• Accepts .pdf input resumes from users
• Returns structured analysis as JSON:
  • Used by frontend to render charts and feedback

---

🔬 Core Features

---

🔍 Tech Stack

---

⚙️ How It Works

1. User Uploads Resumes

   • .pdf files are uploaded to Flask server

2. Text Extraction

   • extract_text_from_pdf() reads content using pdfminer

3. Information Extraction

   • extract_name(), extract_emails(), extract_phone_number()
   • Skill extraction using a predefined keyword list
   • Sections like education/experience parsed line-by-line

4. ATS Scoring

   • Points assigned per section (education, experience, tech skills, etc.)
   • Scoring logic in calculate_ats_score() with feedback

5. Detailed Analysis

   • Word cloud data from generate_word_cloud_data()
   • Skill count data from calculate_skill_distribution()
   • Summary stats like profile completeness, education level

6. Visualization

   • Resume preview iframe
   • Skill tags shown as chips
   • ATS score in circular chart
   • Word cloud rendered with D3
   • Modal opens with full breakdown

---

📊 ATS Score Example

---

🔗 External Dependencies

• NLP Model: spaCy en_core_web_sm
• PDF Text Extraction: pdfminer.six
• Visualization:
  • Chart.js
  • D3.js + d3-cloud

---

📷 Screenshots

---

💡 Future Improvements

• Add job description matching for ranking candidates
• Enable export of results to Excel or PDF
• Integrate a login system for HR teams
• Add support for DOCX resumes

---

📜 License

This project is for educational and academic purposes.

© 2024 Akshay Karthick – Final Year CSE Project