Digital Behavior, Mental Health & Academic Outcomes: A Comprehensive Data Analytics Study on Teen Social Media Usage

By abhishek.verma75000 · May 20, 2026

Datastam conducted a multi-dimensional analytical study examining the relationship between social media behavior and key mental health indicators among…

The analysis reveals that addiction level has very little impact on academic performance and only a modest effect on sleep hours. Two charts were generated — a dual-axis line chart showing trends across addiction levels, and a box plot showing the distribution of academic performance by addiction level. The correlations are nearly zero, suggesting no strong relationship between addiction level and either metric.

The analysis produced two visualizations comparing stress and anxiety levels across different platform usage patterns. A grouped bar chart shows average stress (red) and anxiety (blue) levels side by side for each platform, while a heatmap provides a color-coded overview of how each platform compares on both metrics. These charts let you quickly identify which platforms are associated with the highest mental health burden.

The analysis examined how screen time before sleep relates to sleep duration and stress levels. Two visualizations were generated: a combined bar and line chart showing average sleep hours and stress levels across screen time bins, and a scatter plot showing individual data points colored by stress level. The correlations found were extremely weak — essentially no meaningful relationship between screen time before sleep and either sleep hours (r = 0.010) or stress levels (r = -0.009).

There is a weak positive correlation (r = 0.175) between daily social media usage and depression rates among teens in this dataset. This means that as social media usage increases, depression rates tend to rise slightly, but the relationship is not strong. Two visualizations were generated: a bar chart showing depression rates across usage brackets (0-2h, 2-4h, 4-6h, 6-8h, 8h+), and a bubble scatter plot showing depression rate vs. hours of usage with a trend line. The overall depression rate in the dataset is 2.58%.

The analysis successfully produced heatmaps and data tables showing depression rates and average anxiety levels broken down by age and gender. Two visualizations were generated: a red heatmap showing depression rate (%) by age and gender, and a blue heatmap showing average anxiety levels by age and gender. These charts make it easy to spot which age-gender combinations carry the highest risk at a glance.

Based on the analysis, higher physical activity does NOT meaningfully reduce stress, anxiety, or depression levels in this dataset. The correlations are extremely weak — nearly zero — suggesting physical activity level has little to no measurable impact on these mental health indicators here.

The correlation matrix for all 9 numeric variables has been computed and visualized. A heatmap with annotated correlation coefficients shows the relationships between daily social media hours, sleep hours, screen time before sleep, academic performance, physical activity, stress level, anxiety level, addiction level, and depression label. A second bar chart highlights the top 5 strongest positive and top 5 strongest negative correlations side by side, making it easy to spot the most impactful variable pairs at a glance.

The grouped bar chart and statistics tables show the breakdown of stress and anxiety levels across low, medium, and high social interaction groups. Surprisingly, teens with low social interaction do NOT show significantly higher stress or anxiety — the differences across all three groups are very small (less than 0.2 points on a 1–10 scale).

The IQR-based outlier analysis was completed across all 5 numeric variables: daily social media hours, sleep hours, screen time before sleep, addiction level, and stress level. Remarkably, no outliers were detected in any of the variables using the standard 1.5x IQR rule. Box plots and a bar chart were generated to visualize the distributions and outlier counts. The data appears to be very well-behaved with no extreme values falling outside the expected ranges.

The analysis ranked 3 platform usage categories — Instagram, TikTok, and 'Both' (using both platforms) — by their overall harm to mental health across addiction, stress, anxiety, sleep, and academic performance. A ranked comparison table and bar chart were generated showing each platform's harm score. Instagram came out as the most harmful platform with a harm score of 3.415, followed by TikTok (2.739), and 'Both' platforms combined (2.000) as the least harmful. A radar/spider chart was also produced showing each platform's mental health profile across all five dimensions.