Logo Lanfrica

theaderemifelix/job-scam-detector

Domaine:

socioeconomic

Type de record:

software
Créateur:
the
Hôte:
A machine learning and data visualization project to identify and reduce job scams in Nigeria # 🧠 Job Scam Detection Using Machine Learning & Power BI This project was developed as part of the **3MTT Knowledge Showcase**. It tackles the growing concern of **fake job postings** by combining **machine learning** with **data visualization** to help identify and reduce job scams in Nigeria's labor market. The project involves: - Cleaning and analyzing real-world job listing data - Building a machine learning model to predict scam likelihood - Designing a Power BI dashboard to visualize patterns and risks - Creating a Streamlit app for real-time scam prediction --- ## 📊 Dashboard Summary The Power BI dashboard features: - Real vs Fake Job Distribution - Most Common Job Titles - Job Scam Trends by Category, Location, and Employment Type - ML Prediction Accuracy - Interactive slicers and filters --- ## 🛠️ Tools & Technologies - **Python** (Pandas, Scikit-learn, Joblib) - **Streamlit** – for the AI web app - **Power BI** – for visual storytelling - **Jupyter Notebook** – for model training - **Power Query** – for data transformation - **GitHub** – for version control and sharing --- ## 🧪 Machine Learning Model - **Type:** Logistic Regression - **Accuracy:** 97% - **Features Used:** Job title, description, requirements, company info, salary, location, etc. The model was trained using a cleaned dataset of 202 Nigerian job postings, labeled as either **real (0)** or **fake (1)**. --- ## 🚀 How to Use This Project ### 🔹 Clone the Repository ```bash git clone github.com cd job-scam-detector ``` ### 🔹 Install Requirements ```bash pip install -r requirements.txt ``` ### 🔹 Run the Streamlit App ```bash streamlit run streamlit_app.py ``` --- ## 📁 Project Structure ```plaintext ├── job listings cleaned.csv # Cleaned dataset ├── job_model.pkl # Trained ML model ├── vectorizer.pkl # Text transformer ├── streamlit_app.py # Web app script ├── Notebook.i …