Teen_Mental_Health_Analysis

Exploratory Data Analysis Report

Objective

This report summarizes exploratory findings from the teen mental health dataset. The goal is to understand basic distributions and relationships between lifestyle, social media usage, sleep, academic performance, stress, anxiety, addiction level, and depression label.

Dataset Summary

Metric	Value
Rows	1,200
Columns	13
Missing values	0
Duplicate rows	0
Age range	13 to 19

Numeric Summary

Variable	Mean	Std	Min	Median	Max
age	15.93	2.02	13.00	16.00	19.00
daily_social_media_hours	4.54	2.03	1.00	4.50	8.00
sleep_hours	6.45	1.44	4.00	6.50	9.00
screen_time_before_sleep	1.74	0.72	0.50	1.80	3.00
academic_performance	2.99	0.58	2.00	2.99	4.00
physical_activity	1.01	0.58	0.00	1.00	2.00
stress_level	5.45	2.90	1.00	5.00	10.00
anxiety_level	5.64	2.86	1.00	6.00	10.00
addiction_level	5.57	2.83	1.00	6.00	10.00

Depression Label Distribution

Depression Label	Count	Percent
0	1,169	97.42%
1	31	2.58%

The dataset is highly imbalanced. Only 31 records belong to label 1, so any comparison involving this group should be interpreted carefully.

Sleep Insights

Average sleep duration is approximately 6.45 hours.

Sleep Segment	Percent
Under 6 hours	40.00%
6 to 8 hours	42.83%
Under 8 hours	80.17%

Interpretation:

• Most teens in the dataset sleep below 8 hours.
• The central range from the 25th to 75th percentile is approximately 5.2 to 7.6 hours.

Depression Label Group Comparison

Average Daily Social Media Hours

Depression Label	Average Daily Social Media Hours
0	4.48
1	6.72

Average Sleep Hours

Depression Label	Average Sleep Hours
0	6.49
1	4.76

Average Stress, Anxiety, And Addiction Scores

Depression Label	Stress	Anxiety	Addiction
0	5.37	5.56	5.57
1	8.48	8.61	5.32

Interpretation:

• Label 1 has higher average daily social media use than label 0.
• Label 1 has lower average sleep hours than label 0.
• Label 1 has higher average stress and anxiety scores than label 0.
• Addiction score is similar across labels in this dataset.

Because the label 1 group has only 31 records, these findings should be treated as descriptive signals, not final conclusions.

Academic Performance Insights

Academic performance has:

• Mean: 2.99
• Median: 2.99
• Min: 2.00
• Max: 4.00
• 75th percentile: 3.48

Records with academic performance above 3.5:

Metric	Value
Count	285
Percent	23.75%

Gender distribution for academic performance above 3.5:

Gender	Count
female	151
male	134

Age distribution for academic performance above 3.5:

Age	Count
13	52
14	41
15	36
16	32
17	43
18	37
19	44

Correlation Highlights

Selected correlations with depression_label:

Feature	Correlation
daily_social_media_hours	0.175
sleep_hours	-0.191
stress_level	0.170
anxiety_level	0.170
addiction_level	-0.014
academic_performance	0.001

Interpretation:

• Depression label has a small positive correlation with daily social media hours, stress level, and anxiety level.
• Depression label has a small negative correlation with sleep hours.
• Academic performance has almost no linear relationship with depression label in this dataset.
• These correlations are weak and descriptive; they do not prove causation.

Recommendations For Next Analysis

• Use boxplots to compare sleep and social media hours by depression label.
• Use heatmaps or crosstabs to understand stress and anxiety combinations.
• Consider label imbalance before building any predictive model.
• If modeling is added later, evaluate with metrics suitable for imbalanced data, such as precision, recall, F1-score, and ROC-AUC.

Conclusion

The dataset suggests that teens labeled with depression label 1 tend to have lower sleep hours and higher social media usage, stress, and anxiety scores. However, the label 1 group is small, so findings should be treated as early EDA observations rather than definitive conclusions.