Coursera Dataset Exploration

Overview

This project explores a dataset of Coursera learning programs to uncover trends, insights, and actionable takeaways about program popularity, student enrollment, and certificate types. The analysis aims to answer the key business question:

Which type of learning program is most likely to attract and retain students?

The dataset was sourced from Kaggle and includes 6 columns describing 890 Coursera programs.

---

Business Question

• What learning program onboarded would be most successful?

---

Key Insights & Takeaways

1. Professional Certificates attract the most students on average
   Despite being less common, they consistently show higher enrollment numbers than courses or specializations.

2. Professional Certificates are underrepresented
   They have the lowest prevalence, suggesting a potential growth opportunity.

3. Courses dominate in availability, especially for beginners and mixed-level learners
   Courses cater broadly to beginners and mixed-level students, accounting for the largest share of total enrollments.

---

Dataset

• Raw Source: Scraped from Coursera website
• Columns: Program, Organization, Type, Rating, Difficulty, Students
• Observations: 890 learning programs
• Cleaned Data: A preprocessed version of the dataset is available at data/coursera_data_cleaned.csv for direct use in analysis.

---

Cleaning & Preparation

• Renamed columns for readability.
• Converted Students column from string to integer, handling suffixes (k, M).
• Corrected Type column capitalization.
• Checked for duplicates, null values, and outliers.
• Converted appropriate columns to categorical or numeric types for better performance.
• Exported cleaned CSV for reproducibility.

---

Analysis

• Explored ratings across program types and difficulties.
• Examined student enrollment distribution, including medians and outliers.
• Conducted correlation analysis between ratings and enrollment numbers.
• Investigated organization contributions and program counts.
• Aggregated enrollments by program type and difficulty.

Visualizations include:

• Boxplots for ratings and enrollments
• Scatterplots for rating vs. enrollment
• Stacked bar charts for total enrollments by program type and difficulty

---

Limitations

• No timestamp data, preventing temporal analysis.
• No unique student IDs; some students may appear in multiple programs.
• No competitor or industry-specific data.
• Limited to Coursera platform only.

---

Next Steps

• Re-evaluate trends after 6 months.
• Scrape competitor data for comparison.
• Expand analysis to industry categorization and program completion rates.

---

Getting Started

1. Clone the repository
2. Install dependencies:

pip install pandas numpy seaborn matplotlib

3. Run the notebook Coursera_Dataset_Exploration.ipynb in Jupyter or VS Code.
4. Optionally, use the cleaned dataset directly (data/coursera_data_cleaned.csv) to skip the cleaning steps.