🛑 Fake Job Posting Detection

In today's fast-paced digital landscape, online job postings serve as a vital resource for job seekers. However, the alarming rise of fraudulent job listings has created a challenging environment, making it difficult to distinguish between legitimate opportunities and scams.

This project aims to develop an advanced Fake Job Posting Detection System using Python, SQL, Excel, and Machine Learning—specifically leveraging the Random Forest Model. By implementing cutting-edge data science techniques, we strive to enhance the security of online job portals and empower job seekers to navigate the job market safely and effectively.

---

🎯 Project Objectives

✅ Detect fraudulent job postings to protect job seekers from scams.
✅ Provide statistical insights into the characteristics of fraudulent postings.
✅ Develop a machine learning model to classify job postings based on structured and textual features.
✅ Enhance job portal credibility by flagging suspicious job listings.

---

🛠️ Tech Stack

• Python: Data preprocessing, Machine Learning, Visualization
• SQL: Data storage and querying
• Excel: Data exploration and visualization
• Machine Learning: Random Forest, Logistic Regression, Natural Language Processing (NLP)

---

🔄 Project Workflow

1️⃣ Data Preprocessing (Python)

• Load the dataset using pandas.
• Handle missing values and duplicates.
• Convert categorical data into numerical format.
• Text Processing: Tokenization, stopword removal, HTML tag removal, and stemming.
• Feature Engineering: Generate new relevant features such as word count and keyword frequency.

2️⃣ SQL Import & Exploration

• Store structured data (job title, company, location, etc.) in a MySQL database.
• Run SQL queries to extract insights:
  • Total job postings per country and industry.
  • Most common keywords in fraudulent job postings.
  • Percentage of remote vs. non-remote jobs.
  • Detect duplicate job postings.

3️⃣ Data Visualization (Excel)

📊 Visualizations include:

• Experienced vs. Fraud Job Postings
• Job Posting vs. Presence of Logos
• State-wise Fraud Job Postings
• Global Distribution of Job Postings
• Industry-Level Fraud Analysis

4️⃣ Statistical Analysis

• Hypothesis: Fake job postings contain specific buzzwords more frequently.
• Perform word frequency analysis comparing real vs. fake postings.
• Conduct a chi-square test to check the statistical significance of word usage.

5️⃣ Machine Learning Model

🔹 Feature Selection:

• Structured Data: Job type, location, telecommuting status.
• Textual Data: Job description, requirements (processed with TF-IDF).

🔹 Model Training & Comparison:

• Compare models: Logistic Regression vs. Random Forest.
• Random Forest Classification was chosen due to high accuracy and feature importance analysis.

6️⃣ Performance Evaluation

✅ Key Evaluation Metrics:

• Accuracy
• Precision
• Recall
• F1-score (Crucial for fraud detection)
• ROC Curve & AUC Score

---

📌 Conclusion

This project presents a robust Fake Job Posting Detection System, utilizing machine learning and statistical analysis to safeguard job seekers from fraudulent postings. By leveraging Python, SQL, Excel, and NLP, we ensure an efficient and scalable approach to classifying job listings with high accuracy.

🌟 Let's make job searching safer together!

---