# =============================================================
# Title:       BU 350 Final Project — Data Visualization
# Author:      Hali Sarah Parsons
# Date:        May 2026
# Description: Builds on midterm transformations to produce five
#              visualizations (4 chart types + histogram with KDE)
#              exploring digital habits and mental health patterns.
# Dataset:     digital_habits_vs_mental_health.xlsx (Kaggle)
# =============================================================

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import matplotlib.patches as mpatches
import seaborn as sns
from scipy import stats

%matplotlib inline
plt.rcParams['figure.dpi'] = 120
plt.rcParams['font.family'] = 'sans-serif'

print("Libraries loaded.")

Libraries loaded.

# =============================================================
# Load the raw dataset and reapply all five midterm transformations
# in sequence so this notebook is self-contained and reproducible.
# =============================================================

df = pd.read_excel('digital_habits_vs_mental_health.xlsx')

# --- Transformation 1: Rename columns to readable labels ---
df.rename(columns={
    'screen_time_hours':           'Daily Screen Time (hrs)',
    'social_media_platforms_used': 'Platforms Used',
    'hours_on_TikTok':             'TikTok Usage (hrs)',
    'sleep_hours':                 'Sleep Duration (hrs)',
    'stress_level':                'Stress Level (1-10)',
    'mood_score':                  'Mood Score (1-10)'
}, inplace=True)

# --- Transformation 2: Remove duplicate rows ---
df.drop_duplicates(inplace=True)
df.reset_index(drop=True, inplace=True)

# --- Transformation 3: Flag screen time outliers using IQR ---
col = 'Daily Screen Time (hrs)'
Q1 = df[col].quantile(0.25)
Q3 = df[col].quantile(0.75)
IQR = Q3 - Q1
df['Screen Time Outlier'] = (df[col] < Q1 - 1.5 * IQR) | (df[col] > Q3 + 1.5 * IQR)

# --- Transformation 4: Categorize sleep duration ---
def categorize_sleep(hours):
    if hours < 5:   return 'Severely Insufficient'
    elif hours < 6: return 'Insufficient'
    elif hours < 7: return 'Below Recommended'
    elif hours <= 9: return 'Recommended'
    else:           return 'Excessive'

df['Sleep Quality Category'] = df['Sleep Duration (hrs)'].apply(categorize_sleep)

# --- Transformation 5: Standardize decimal precision ---
for c in ['Daily Screen Time (hrs)', 'TikTok Usage (hrs)', 'Sleep Duration (hrs)']:
    df[c] = df[c].round(1)

print(f"Dataset shape after midterm transformations: {df.shape}")
print(f"Columns: {list(df.columns)}")
df.head()

Dataset shape after midterm transformations: (98135, 8)
Columns: ['Daily Screen Time (hrs)', 'Platforms Used', 'TikTok Usage (hrs)', 'Sleep Duration (hrs)', 'Stress Level (1-10)', 'Mood Score (1-10)', 'Screen Time Outlier', 'Sleep Quality Category']

# =============================================================
# Transformation 6: Bin stress level into three interpretable tiers
# Low (1-3), Moderate (4-7), High (8-10) for grouped comparisons
# =============================================================

stress_bins   = [0, 3, 7, 10]
stress_labels = ['Low (1-3)', 'Moderate (4-7)', 'High (8-10)']

df['Stress Category'] = pd.cut(
    df['Stress Level (1-10)'],
    bins=stress_bins,
    labels=stress_labels,
    right=True
)

print("Stress Category distribution:")
print(df['Stress Category'].value_counts().sort_index())

Stress Category distribution:
Stress Category
Low (1-3)          9597
Moderate (4-7)    61809
High (8-10)       26729
Name: count, dtype: int64

# =============================================================
# Transformation 7: Bin daily screen time into named usage tiers
# Provides a labeled categorical variable for proportional charts
# =============================================================

screen_bins   = [0, 3, 6, 9, 13]
screen_labels = ['Light (1-3 hrs)', 'Moderate (3-6 hrs)', 'Heavy (6-9 hrs)', 'Excessive (9+ hrs)']

df['Screen Time Tier'] = pd.cut(
    df['Daily Screen Time (hrs)'],
    bins=screen_bins,
    labels=screen_labels,
    right=True
)

print("Screen Time Tier distribution:")
print(df['Screen Time Tier'].value_counts().sort_index())
print(f"\nFinal dataset shape: {df.shape}")
df.head()

Screen Time Tier distribution:
Screen Time Tier
Light (1-3 hrs)        6745
Moderate (3-6 hrs)    43200
Heavy (6-9 hrs)       41702
Excessive (9+ hrs)     6488
Name: count, dtype: int64

Final dataset shape: (98135, 10)

# =============================================================
# Visualization 1: Seaborn Bar Chart
# Mean Mood Score by Stress Category (Low / Moderate / High)
# Error bars show 95% confidence interval around each group mean.
# Color maps intuitively: green for low stress, red for high stress.
# =============================================================

palette_stress = {'Low (1-3)': '#27AE60', 'Moderate (4-7)': '#E67E22', 'High (8-10)': '#C0392B'}

fig, ax = plt.subplots(figsize=(9, 6))

sns.barplot(
    data=df,
    x='Stress Category',
    y='Mood Score (1-10)',
    palette=palette_stress,
    order=stress_labels,
    errorbar='ci',
    capsize=0.06,
    ax=ax
)

# Annotate mean value above each bar
for patch, label in zip(ax.patches, stress_labels):
    mean_val = df[df['Stress Category'] == label]['Mood Score (1-10)'].mean()
    ax.text(
        patch.get_x() + patch.get_width() / 2,
        patch.get_height() + 0.05,
        f'{mean_val:.2f}',
        ha='center', va='bottom',
        fontsize=12, fontweight='bold'
    )

ax.set_title("Visualization 1 — Mean Mood Score by Stress Category", fontsize=14, fontweight='bold')
ax.set_xlabel("Stress Category", fontsize=12)
ax.set_ylabel("Mean Mood Score (1-10)", fontsize=12)
ax.set_ylim(0, 11)
ax.yaxis.grid(True, linestyle='--', alpha=0.5)
ax.set_axisbelow(True)
sns.despine()
plt.tight_layout()
plt.savefig('viz1_bar.png', dpi=120, bbox_inches='tight')
plt.show()
print("Visualization 1 rendered.")

/tmp/ipykernel_562/1555461023.py:12: FutureWarning: 

Passing `palette` without assigning `hue` is deprecated and will be removed in v0.14.0. Assign the `x` variable to `hue` and set `legend=False` for the same effect.

  sns.barplot(

Visualization 1 rendered.

# =============================================================
# Visualization 2: Seaborn Box Plot
# Sleep Duration distribution across Sleep Quality Categories.
# The IQR-based whiskers match the outlier rule from the midterm.
# Ordered from least to most sleep for logical left-to-right reading.
# =============================================================

sleep_order = ['Severely Insufficient', 'Insufficient', 'Below Recommended', 'Recommended', 'Excessive']
palette_sleep = sns.color_palette("Blues", n_colors=5)

fig, ax = plt.subplots(figsize=(11, 6))

sns.boxplot(
    data=df,
    x='Sleep Quality Category',
    y='Sleep Duration (hrs)',
    order=sleep_order,
    palette=palette_sleep,
    linewidth=1.4,
    flierprops={'marker': 'o', 'markersize': 2, 'alpha': 0.3},
    ax=ax
)

ax.set_title("Visualization 2 — Sleep Duration by Sleep Quality Category", fontsize=14, fontweight='bold')
ax.set_xlabel("Sleep Quality Category", fontsize=12)
ax.set_ylabel("Sleep Duration (hours/night)", fontsize=12)
ax.tick_params(axis='x', labelsize=10)
ax.yaxis.grid(True, linestyle='--', alpha=0.5)
ax.set_axisbelow(True)
sns.despine()
plt.tight_layout()
plt.savefig('viz2_boxplot.png', dpi=120, bbox_inches='tight')
plt.show()
print("Visualization 2 rendered.")

/tmp/ipykernel_562/1363800933.py:13: FutureWarning: 

Passing `palette` without assigning `hue` is deprecated and will be removed in v0.14.0. Assign the `x` variable to `hue` and set `legend=False` for the same effect.

  sns.boxplot(

Visualization 2 rendered.

# =============================================================
# Visualization 3: Seaborn Scatter Plot with Regression Lines
# Screen time vs. Mood Score, colored by Stress Category.
# alpha=0.15 prevents overplotting at 98,000+ points.
# A lmplot-style overlay shows linear trend per stress group.
# =============================================================

palette_stress_scatter = {
    'Low (1-3)':    '#27AE60',
    'Moderate (4-7)': '#E67E22',
    'High (8-10)':  '#C0392B'
}

fig, ax = plt.subplots(figsize=(11, 7))

for category, color in palette_stress_scatter.items():
    subset = df[df['Stress Category'] == category]
    ax.scatter(
        subset['Daily Screen Time (hrs)'],
        subset['Mood Score (1-10)'],
        alpha=0.08,         # heavy transparency handles 98k+ point overplotting
        s=6,
        color=color,
        label=f'Stress: {category}'
    )
    # Add a linear regression line per group to show directional trend
    m, b, r, p, se = stats.linregress(subset['Daily Screen Time (hrs)'], subset['Mood Score (1-10)'])
    x_line = np.linspace(subset['Daily Screen Time (hrs)'].min(), subset['Daily Screen Time (hrs)'].max(), 100)
    ax.plot(x_line, m * x_line + b, color=color, linewidth=2.2, alpha=0.9)

ax.set_title(
    "Visualization 3 — Daily Screen Time vs. Mood Score by Stress Category",
    fontsize=14, fontweight='bold'
)
ax.set_xlabel("Daily Screen Time (hours)", fontsize=12)
ax.set_ylabel("Mood Score (1-10)", fontsize=12)
ax.legend(title="Stress Category", fontsize=10, title_fontsize=10)
ax.spines['top'].set_visible(False)
ax.spines['right'].set_visible(False)
plt.tight_layout()
plt.savefig('viz3_scatter.png', dpi=120, bbox_inches='tight')
plt.show()
print("Visualization 3 rendered.")

Visualization 3 rendered.

# =============================================================
# Visualization 4: Seaborn Heatmap — Correlation Matrix
# Lower triangle only (upper is redundant). Diagonal masked.
# coolwarm diverging palette: blue = negative, red = positive.
# Annotated with 2-decimal correlation values for exact reading.
# =============================================================

numeric_cols = [
    'Daily Screen Time (hrs)', 'Platforms Used', 'TikTok Usage (hrs)',
    'Sleep Duration (hrs)', 'Stress Level (1-10)', 'Mood Score (1-10)'
]

corr = df[numeric_cols].corr()

# Mask: True = hidden. triu masks upper triangle including diagonal.
mask = np.triu(np.ones_like(corr, dtype=bool))

display_labels = ['Screen Time', 'Platforms Used', 'TikTok Hours',
                  'Sleep Duration', 'Stress Level', 'Mood Score']

fig, ax = plt.subplots(figsize=(9, 7))

sns.heatmap(
    corr,
    mask=mask,
    annot=True,
    fmt='.2f',
    cmap='coolwarm',
    center=0,
    vmin=-1, vmax=1,
    linewidths=0.5,
    linecolor='white',
    xticklabels=display_labels,
    yticklabels=display_labels,
    ax=ax
)

ax.set_title(
    "Visualization 4 — Pearson Correlation Matrix: Digital Habits & Mental Health",
    fontsize=13, fontweight='bold', pad=15
)
plt.xticks(rotation=30, ha='right', fontsize=10)
plt.yticks(rotation=0, fontsize=10)
plt.tight_layout()
plt.savefig('viz4_heatmap.png', dpi=120, bbox_inches='tight')
plt.show()
print("Visualization 4 rendered.")

Visualization 4 rendered.

# =============================================================
# Visualization 5: Histogram + Normal Distribution Curve Overlay
# Required chart for the final project submission.
# scipy.stats.norm.fit() estimates mu and sigma from the actual data.
# IQR bounds from Transformation 3 are marked for reference.
# density=True aligns the histogram y-axis with the PDF curve scale.
# =============================================================

screen = df['Daily Screen Time (hrs)'].dropna()

# Fit a normal distribution to the observed screen time data
mu, std = stats.norm.fit(screen)

# Recalculate IQR bounds for annotation
Q1_s = screen.quantile(0.25)
Q3_s = screen.quantile(0.75)
IQR_s = Q3_s - Q1_s
lower_iqr = Q1_s - 1.5 * IQR_s
upper_iqr = Q3_s + 1.5 * IQR_s

fig, ax = plt.subplots(figsize=(11, 6))

# Histogram bars — density=True scales to probability density
ax.hist(
    screen,
    bins=24,
    density=True,
    color='#4C9A8A',
    edgecolor='white',
    linewidth=0.6,
    alpha=0.78,
    label='Observed Distribution'
)

# Fitted normal curve across the observed range
x_range = np.linspace(screen.min(), screen.max(), 400)
pdf_fitted = stats.norm.pdf(x_range, mu, std)
ax.plot(x_range, pdf_fitted, color='#2B6B5E', linewidth=2.8,
        label=f'Fitted Normal Curve  μ={mu:.2f}, σ={std:.2f}')

# Mean and median reference lines
ax.axvline(mu, color='#C0392B', linewidth=1.8, linestyle='--',
           label=f'Mean = {mu:.2f} hrs')
ax.axvline(screen.median(), color='#E67E22', linewidth=1.8, linestyle=':',
           label=f'Median = {screen.median():.2f} hrs')

# IQR outlier bounds — show where the midterm flagging threshold sits
ax.axvline(lower_iqr, color='#7F8C8D', linewidth=1.2, linestyle='-.',
           alpha=0.7, label=f'IQR Lower = {lower_iqr:.1f} hrs')
ax.axvline(upper_iqr, color='#7F8C8D', linewidth=1.2, linestyle='-.',
           alpha=0.7, label=f'IQR Upper = {upper_iqr:.1f} hrs')

ax.set_title(
    "Visualization 5 — Distribution of Daily Screen Time with Fitted Normal Curve\n"
    "(IQR outlier bounds from midterm Transformation 3 marked in grey)",
    fontsize=13, fontweight='bold'
)
ax.set_xlabel("Daily Screen Time (hours/day)", fontsize=12)
ax.set_ylabel("Probability Density", fontsize=12)
ax.legend(fontsize=10, loc='upper right')
ax.spines['top'].set_visible(False)
ax.spines['right'].set_visible(False)
plt.tight_layout()
plt.savefig('viz5_hist_kde.png', dpi=120, bbox_inches='tight')
plt.show()
print(f"Visualization 5 rendered.  μ={mu:.2f}, σ={std:.2f}")
print(f"IQR outlier bounds: [{lower_iqr:.2f}, {upper_iqr:.2f}]")

Visualization 5 rendered.  μ=6.02, σ=1.99
IQR outlier bounds: [0.65, 11.45]

BU 350 — Final Project: Data Visualization¶

Digital Habits vs. Mental Health Dataset¶

Step 1 — Load Dataset and Reapply Midterm Transformations¶

Step 2 — Additional Transformations for Visualization¶

Visualization 1 — Bar Chart: Mean Mood Score by Stress Category¶

Visualization 2 — Box Plot: Sleep Duration by Sleep Quality Category¶

Visualization 3 — Scatter Plot: Screen Time vs. Mood Score (by Stress Category)¶

Visualization 4 — Heatmap: Correlation Matrix Across All Numeric Variables¶

Visualization 5 — Histogram with Distribution Curve: Daily Screen Time¶

Final Paper¶

References¶

	Daily Screen Time (hrs)	Platforms Used	TikTok Usage (hrs)	Sleep Duration (hrs)	Stress Level (1-10)	Mood Score (1-10)	Screen Time Outlier	Sleep Quality Category
0	12.0	4	7.1	10.0	7	9	True	Excessive
1	11.3	1	6.5	10.0	9	10	False	Excessive
2	10.8	2	6.4	10.0	7	10	False	Excessive
3	10.8	5	6.2	10.0	7	10	False	Excessive
4	11.3	4	6.1	10.0	5	8	False	Excessive