Lab-PDF-To-Dataset

Developed by Paksa IT Solutions

Overview

Lab-PDF-To-Dataset is an intelligent data extraction system that automatically converts medical laboratory test reports (PDF and Word documents) into structured, machine-readable CSV and Excel datasets. This tool eliminates the tedious manual data entry process and provides clean, ready-to-use datasets for analysis and research.

Problem It Solves

Medical lab reports are typically stored as unstructured PDF or Word documents, making it extremely difficult to:

• Extract data for analysis
• Build datasets for machine learning models
• Conduct statistical research
• Perform comparative studies across multiple patients

Our system automates this entire process, converting hundreds of lab reports into structured datasets in minutes.

Key Features

• Multi-Format Support: Extract data from PDF (text-based and scanned) and Word documents (.doc, .docx)
• OCR Capabilities: Automatically detects scanned PDFs and uses OCR (Tesseract) to extract text.
• Multiple Test Types: Supports CBC (Complete Blood Count), LFT (Liver Function Test), RFT (Renal Function Test), and TFT (Thyroid Function Test)
• Dual Output: Generates both CSV and Excel (.xlsx) datasets.
• Batch Processing: Upload ZIP files containing multiple reports for bulk processing
• Robust Error Handling: Skips problematic files without stopping the batch, logging errors to a dedicated CSV report.
• Web Interface: User-friendly React-based frontend with progress tracking and direct downloads.
• Accurate Extraction: Smart regex patterns to extract actual test results, avoiding ranges and dates.

Benefits for Students

1. Academic Research Projects

• Quickly build datasets for medical informatics projects
• Analyze health trends across patient demographics
• Complete assignments requiring real-world medical data

2. Learning Data Processing

• Understand how unstructured data (including scanned images) is converted to structured formats
• Learn about regex patterns, OCR, and text extraction techniques
• Study the architecture of full-stack applications (Flask + React)

3. Final Year Projects

• Use generated datasets for machine learning projects
• Build predictive models for disease diagnosis
• Create health analytics dashboards

Benefits for Data Scientists

1. Rapid Dataset Creation

• Convert thousands of lab reports into analysis-ready datasets in minutes
• eliminate weeks of manual data entry work
• Focus on analysis rather than data collection

2. High-Quality Data

• Excel Export: Get data in a format ready for immediate analysis in Excel, pandas, or other tools.
• Error Logging: detailed Processing_Errors.csv helps identify data quality issues source files.
• Consistent Formatting: Standardized columns across all extracted files.

3. Machine Learning Applications

• Build predictive models for disease detection
• Train classification algorithms for abnormal test results
• Develop patient risk assessment systems

Installation

Prerequisites

1. Python 3.8+
2. Node.js 14+
3. Tesseract OCR (For scanned PDFs):
   • Windows: Download Installer (Add to PATH)
   • Linux: sudo apt install tesseract-ocr
   • Mac: brew install tesseract
4. Poppler (For PDF-to-Image conversion):
   • Windows: Download Release (Add bin folder to PATH)
   • Linux: sudo apt install poppler-utils
   • Mac: brew install poppler

Backend Setup

cd lab_pdf_to_dataset
pip install -r requirements.txt
python app.py

Frontend Setup

cd frontend
npm install
npm run dev

Usage

1. Start the Backend: Run python app.py (Flask server on port 5000)
2. Start the Frontend: Run npm run dev (React app on port 5173)
3. Upload Files:
   • Single or Multiple PDF/Word files (Text or Scanned)
   • ZIP archive containing multiple reports
4. Select Test Types: Check boxes for CBC, LFT, RFT, or TFT.
5. Download Results:
   • Download CSV or Excel datasets directly from the UI.
   • Download Error Log if any files failed.

Output Format

CBC Dataset Columns

Name, Age, Gender, HB, RBC, HCT, MCV, MCH, MCHC, Platelets, WBC, Neutrophils, Lymphocytes, Monocytes, Eosinophils, ESR, Source_PDF

LFT Dataset Columns

Name, Age, Gender, Total Bilirubin, Direct Bilirubin, Indirect Bilirubin, ALT, AST, ALP, Albumin, Total Protein, Source_PDF

RFT Dataset Columns

Name, Age, Gender, Urea, BUN, Creatinine, GFR, Uric Acid, Source_PDF

TFT Dataset Columns

Name, Age, Gender, T3, T4, TSH, Free T3, Free T4, Source_PDF

---

Developed by Paksa IT Solutions
© 2026 Paksa IT Solutions. All rights reserved.