Power BI AI CV Analysis for Recruitment: Automating Candidate Matching with OpenAI

Obtain clear, explicit consent from individuals before processing their data.

Limit data usage strictly to the recruitment process, avoiding any secondary purposes.

Remove or mask personally identifiable information (PII) before processing CVs to minimize privacy risks.

This includes names, contact details, and other sensitive identifiers.

Familiarize yourself with OpenAI’s usage policies to avoid transmitting unnecessary sensitive information.

Ensure secure transmission of data to and from the API.

Regularly review AI outputs for potential bias or unfair patterns in candidate evaluations.

Provide additional checks to ensure all candidates are assessed equitably.

Use encryption to protect data during storage and transmission.

Clearly inform candidates about how AI is used in the recruitment process and provide options to withdraw their data if needed.

graph TD
    subgraph "VS Code - Jupyter"
        A[Start]
        B[Extract text from CV PDFs]
        A --> B
        B --> C[OCR_Results.json]
        F[Extract text from Position PDF]
        A --> F
        F --> G[OCR_Position.json]
        
        E[LLM_Normalized_CV.json]
        I[LLM_Position.json]
        K[LLM_Analysis.json]
        
        subgraph "OpenAI"
            D[Summarize CVs]
            H[Normalize Position Description]
            J[Evaluate Candidates]
        end
        
        C --> D
        D --> E
        G --> H
        H --> I
        E --> J
        I --> J
        J --> K
    end
    
    E --> L[Analyze with Power BI]
    I --> L
    K --> L
    
    %% Define styles
    classDef openai color:#FFFFFF, fill:#FF0000,stroke:#FF0000,stroke-width:2px;
    classDef vscode color:#FFFFFF, fill:#008000,stroke:#008000,stroke-width:2px;
    classDef powerbi color:#000000, fill:#FFD700,stroke:#FFD700,stroke-width:2px;

    %% Apply styles
    class D,H,J openai
    class A,B,C,E,F,G,I,K vscode
    class L powerbi

Download Python from the official website and follow the installation instructions for your operating system.

VS Code Official Download Page: Download Visual Studio Code

Setup Guide: Microsoft: Set up VS Code - Official setup overview from Microsoft.

Git for Windows: Download Git

This will enable you to clone the repository and manage version control.

If you’re unfamiliar with Git, these steps will guide you through cloning the repository using VS Code:

1. Open VS Code
2. Open the Command Palette by pressing Ctrl+Shift+P
3. Type “Git: Clone” and select it
4. Enter the Repository URL: https://github.com/OscarValerock/BIBB-PBI-CV-AI-Analysis.git
5. Choose a Local Directory: Select a folder on your computer to store the project
6. Open the Repository: VS Code will prompt you to open the repository once cloned. Click Open

Once you’ve cloned the repository and have your project open in Visual Studio Code, it’s best practice to create a virtual environment for your project. This isolates the required Python packages, making your setup more stable and organized.

Steps to create the virtual environment:

• In the Command Palette (Ctrl+Shift+P), type Python: Create Environment and select Create Environment
• VS Code will prompt you to select a folder. Select your project folder (the cloned repository) and choose venv as the virtual environment type
• Accept to install the packages from requirements.txt

Create a file named Constants.py in your project’s root directory. This file will store your OpenAI API key.

OpenAIKey = "your-openai-api-key"

Important: Replace "your-openai-api-key" with your actual OpenAI API key. Keep this file secure and avoid sharing it publicly.

This script uses PyMuPDF to extract text from all PDF files in the specified CV folder. Each CV’s text is saved in a JSON file, where each entry contains the filename and the extracted content. This JSON file will later serve as input for OpenAI processing.

This structure outlines how the CV data will be organized after processing with OpenAI. It includes sections for:

• Personal information
• Education
• Work experience
• Skills
• Certifications
• Languages
• Systems knowledge

By providing a consistent structure, it enables clear and comparable information across different CVs.

Using OpenAI’s GPT model, this script takes the raw text from each CV and organizes it into the predefined JSON format. The script processes each CV’s content individually, sends it to OpenAI, and then saves the structured output in a new JSON file for further analysis.

Like CV processing, this script extracts text from the job position description PDF and organizes it into a structured JSON format. This structure includes fields for:

• Job title
• Department
• Responsibilities
• Qualifications
• Required skills

This ensures a comparable format for evaluation against CVs.

This final script compares each candidate’s normalized CV with the job description, assessing the match across dimensions like:

• Education
• Experience
• Skills
• Overall fit

It uses OpenAI to evaluate and output the results in JSON format, which can be loaded into Power BI for visualization and further analysis. The script provides a “score” for each dimension to quantify the match.