Analyzing Students' Mental Health in MySQL (Advanced)

1. Project Overview & Objectives

This project builds upon foundational SQL skills to conduct an in-depth analysis of hypothetical international university student mental health data using MySQL. It investigates the impact of length of stay on various mental health diagnostic scores, alongside demographic factors.

• Core Objective: To identify trends and correlations between students' duration at the university and their self-reported anxiety, depression, and social connectedness scores, specifically focusing on international students.
• Advanced Objectives & Skills Demonstrated:
  • Advanced Data Modeling: Designing and implementing a normalized relational database schema with multiple interconnected tables (students, demographics, surveys).
  • Complex SQL Queries: Utilizing JOIN operations, advanced date functions (TIMESTAMPDIFF), Common Table Expressions (CTEs), Window Functions (AVG() OVER (...)), and Conditional Aggregation (CASE within AVG()) for sophisticated data extraction and analysis.
  • Data Visualization: Integrating MySQL data with Python (Pandas, Matplotlib, Seaborn) to create compelling visual insights and trends.
  • Performance Awareness: Conceptual understanding of database indexing for optimization.

---

2. Data Model

The database is designed with three normalized tables to manage student, demographic, and survey data efficiently.

students Table

Stores core student information.

Column Name	Data Type	Description
student_id	INT	Unique identifier (Primary Key, Auto-Increment)
is_international	BOOLEAN	TRUE if international, FALSE if domestic
enrollment_date	DATE	Date student enrolled (used for length_of_stay)

demographics Table

Stores static demographic information, linked to students.

Column Name	Data Type	Description
student_id	INT	Primary Key and Foreign Key referencing students(student_id)
age	INT	Student's age
gender	VARCHAR(10)	Student's gender (e.g., 'Male', 'Female', 'Non-binary')
program_of_study	VARCHAR(100)	Academic program/major
Constraint		FOREIGN KEY (student_id) REFERENCES students(student_id) (CASCADE)

surveys Table

Stores individual survey responses, allowing multiple entries per student over time.

Column Name	Data Type	Description
survey_id	INT	Unique identifier for each survey (Primary Key, Auto-Increment)
student_id	INT	Foreign Key referencing students(student_id)
survey_date	DATE	Date the survey was taken
anxiety_score	INT	Diagnostic score for anxiety (0-100, higher = more anxiety)
depression_score	INT	Diagnostic score for depression (0-100, higher = more depression)
social_connectedness_score	INT	Score for social connectedness (0-100, higher = more connected)
Constraint		FOREIGN KEY (student_id) REFERENCES students(student_id) (CASCADE)

---

3. Key SQL Queries & Python Integration

All SQL queries are executed against a MySQL database. The primary analysis is then fed into a Python script for visualization.

Database and Table Creation (SQL)

-- Drop existing tables (if any) to ensure clean setup
DROP TABLE IF EXISTS surveys;
DROP TABLE IF EXISTS demographics;
DROP TABLE IF EXISTS students;

-- Create students table
CREATE TABLE students (
    student_id INT PRIMARY KEY AUTO_INCREMENT,
    is_international BOOLEAN,
    enrollment_date DATE
);

-- Create demographics table with Foreign Key to students
CREATE TABLE demographics (
    student_id INT PRIMARY KEY, -- FK to students
    age INT,
    gender VARCHAR(10),
    program_of_study VARCHAR(100), -- From 'What is your course?'
    CONSTRAINT fk_student_demographics FOREIGN KEY (student_id) REFERENCES students(student_id)
        ON DELETE CASCADE ON UPDATE CASCADE
);

-- Create surveys table with Foreign Key to students and new fields
CREATE TABLE surveys (
    survey_id INT PRIMARY KEY AUTO_INCREMENT,
    student_id INT, -- FK to students
    survey_date DATETIME, -- From 'Timestamp', now DATETIME
    year_of_study INT, -- From 'Your current year of Study'
    cypa_range VARCHAR(50), -- From 'What is your CGPA?'
    marital_status VARCHAR(20), -- From 'Marital status'
    anxiety_score INT,        -- 'Yes'/'No' converted to 1/0
    depression_score INT,     -- 'Yes'/'No' converted to 1/0
    panic_attack_score INT,   -- 'Yes'/'No' converted to 1/0 (NEW)
    sought_treatment INT,     -- 'Yes'/'No' converted to 1/0 (NEW)
    CONSTRAINT fk_student_surveys FOREIGN KEY (student_id) REFERENCES students(student_id)
        ON DELETE CASCADE ON UPDATE CASCADE
);

Sample Data Insertion (SQL)

Sample data for students, demographics, and surveys tables is inserted to populate the database and enable comprehensive analysis. (Full INSERT statements are omitted for brevity in this README but are available in the project's SQL files, e.g., insert_data.sql.)

Core Advanced Analysis Query (SQL using CTEs, JOINS, Date Functions, Conditional Aggregation)

This query pulls data from all three tables, dynamically calculates length_of_stay_months, filters for international students, and aggregates average mental health scores, including a gender-based breakdown, grouped by length of stay.

USE student_mental_health;

WITH StudentSurveyData AS (
    SELECT
        s.student_id,
        d.age,
        d.gender,
        d.program_of_study,
        su.survey_date,
        su.year_of_study,
        su.cypa_range, -- Corrected column name from 'cgpa_range' in previous iterations
        su.marital_status,
        su.anxiety_score,
        su.depression_score,
        su.panic_attack_score,
        su.sought_treatment
    FROM
        students s
    INNER JOIN
        demographics d ON s.student_id = d.student_id
    INNER JOIN
        surveys su ON s.student_id = su.student_id
)
SELECT
    year_of_study,
    COUNT(DISTINCT student_id) AS number_of_students,
    AVG(anxiety_score) AS average_anxiety_score,
    AVG(depression_score) AS average_depression_score,
    AVG(panic_attack_score) AS average_panic_attack_score,
    -- AVG(social_connectedness_score) AS average_social_connectedness_score, -- Removed as this column is not in the dataset
    AVG(CASE WHEN gender = 'Male' THEN anxiety_score END) AS avg_anxiety_male,
    AVG(CASE WHEN gender = 'Female' THEN anxiety_score END) AS avg_anxiety_female,
    AVG(CASE WHEN sought_treatment = 1 THEN anxiety_score END) AS avg_anxiety_sought_treatment,
    AVG(CASE WHEN sought_treatment = 0 THEN anxiety_score END) AS avg_anxiety_not_sought_treatment
FROM
    StudentSurveyData
GROUP BY
    year_of_study
ORDER BY
    year_of_study;

Data Visualization (Python)

A Python script (visualize_data.py) connects to the MySQL database, executes the core analysis query, loads the results into a Pandas DataFrame, and generates compelling visualizations using Matplotlib and Seaborn. These plots are saved as PNG files and displayed below:

Visualizations

Mental Health Scores by Year

Anxiety by Gender and Year

Anxiety by Treatment and Year

Mental Health Scores by Stay

---

4. Potential Challenges & Learning

• Database Normalization: Designing and implementing a multi-table schema with correct primary and foreign key relationships for data integrity.
• Complex SQL: Mastering JOIN clauses, dynamic date calculations (TIMESTAMPDIFF), and advanced aggregation techniques (CASE statements within AVG(), CTEs, and Window Functions) for nuanced analysis.
• Data Integration & Visualization: Connecting MySQL to Python, querying data programmatically, and using data science libraries (Pandas, Matplotlib, Seaborn) to transform raw data into insightful visualizations.
• Data Interpretation: Drawing meaningful conclusions from aggregated and visualized data, especially with synthetic datasets.

---

5. Real-World Considerations & Future Enhancements

• Data Privacy: In a real-world scenario, student mental health data is highly sensitive. Strict anonymization protocols and adherence to privacy regulations (e.g., GDPR, FERPA) would be paramount.
• Limitations of Synthetic Data: The findings in this project are based on a small, synthetic dataset. Real-world insights would require a larger, validated, and ethically sourced dataset.
• Scalability: For much larger datasets, optimizing queries with appropriate indexing (e.g., on is_international, survey_date, enrollment_date) and analyzing query performance using EXPLAIN would be critical.

Future Enhancements:

• Predictive Modeling: Use Python's machine learning libraries (e.g., scikit-learn) to build models that predict mental health risk based on demographic factors and length of stay.
• Interactive Dashboard: Create an interactive web dashboard (e.g., using Dash or Streamlit in Python, or a BI tool like Tableau/Power BI) to allow users to explore the data dynamically.
• Expanded Data: Incorporate more granular data, such as academic performance, financial aid status, or participation in campus activities, to enrich the analysis.
• API Development: Build a simple API using a framework like Flask or FastAPI in Python to expose the analytical results, allowing other applications to consume this data.