Netflix Attention Span Analysis

This analysis explores viewing patterns and attention span metrics based on Netflix watching history. The visualizations below provide insights into viewing habits, binge-watching patterns, and attention metrics over time.

Data Analysis Techniques

Time-Based Analysis

Examining viewing patterns over different periods (daily, weekly, monthly)

Session Analysis

Defining viewing sessions and analyzing duration/frequency

Content Analysis

Categorizing watched content by genre and type

Attention Metrics

Quantifying attention span through various metrics

Viewing Patterns Analysis

Daily Viewing Patterns

Shows the average viewing duration for each day of the week. This helps identify which days tend to have longer or shorter viewing sessions.

Monthly Viewing Trends

Displays how viewing duration changes across months, helping identify seasonal patterns or long-term trends in viewing habits.

Viewing Pattern Over Time

Shows individual viewing sessions and a 7-day rolling average, highlighting both daily variations and overall trends in viewing duration.

Yearly Analysis

Compares yearly viewing statistics, including average duration and number of sessions per year, to identify long-term changes in viewing habits.

Binge Watching Distribution

Shows the distribution of viewing session durations, categorized by length, to identify binge-watching patterns.

Attention Metrics

Displays key metrics related to viewing attention, including completion rate, session consistency, and overall attention score.

Machine Learning Analysis

Future Viewing Predictions

Our machine learning model achieves 82.0% accuracy in predicting viewing patterns, with a confidence interval of 95.0%. The model analyzes historical viewing data to forecast future trends, taking into account:

  • Daily viewing patterns and peak hours
  • Weekly viewing cycles and weekend effects
  • Seasonal variations throughout the year
  • Long-term behavioral trends

Model Accuracy

82.0%

Overall prediction accuracy score

Confidence Level

95.0%

Prediction confidence interval

The graph below shows both historical data and predicted future viewing patterns, with shaded areas representing the confidence intervals of our predictions.

+27% Weekend Viewing
-12% Morning Hours
+15% Evening Peak

Key Findings Summary

📊

Stable viewing patterns over years despite visual trends

⏰

Peak engagement during 8 PM - 11 PM

📺

30% of sessions are binge-watching

📈

85% average episode completion rate

Key Findings and Statistical Analysis

Trend Analysis

While initial observations might suggest a decrease in average viewing duration over the years, our rigorous statistical analysis tested the following hypotheses:

Alternative Hypothesis (H₁)

The average viewing session duration has decreased over time, indicating a reduction in attention span.

Null Hypothesis (H₀)

There is no significant change in the average viewing session duration over time, indicating stable attention spans.

While initial observations might suggest a decrease in average viewing duration over the years, our rigorous statistical analysis tells a different story:

  • Visual Trend: The raw data shows an apparent decrease in viewing duration
  • Statistical Validation: We conducted a Mann-Kendall trend test to verify this observation:
    • Correlation coefficient: -0.6667
    • p-value: 0.3333 (greater than significance level of 0.05)
    • Result: Failed to reject the null hypothesis
  • Conclusion: Despite the visual appearance, there is no statistically significant trend in viewing duration over the years
Conclusion:

Despite the visual appearance, there is no statistically significant trend in viewing duration over the years. This concluded by a p-value of 0.3333, which fails to reject the null hypothesis.

Primary Observations

  1. Viewing Patterns
    • Peak viewing occurs during evening hours (8 PM - 11 PM)
    • Weekends show 30% higher viewing duration compared to weekdays
    • Seasonal variations exist with higher engagement during winter months
  2. Binge-Watching Behavior
    • Average binge session consists of 3-4 episodes
    • 30% of viewing sessions qualify as binge-watching
    • Weekend binge sessions are typically longer than weekday sessions
  3. Attention Metrics
    • Episode completion rate averages 85%
    • Content engagement is highest for episodes under 40 minutes

Practical Implications

These findings suggest several recommendations:

  • Episode length optimization around 40 minutes may maximize engagement
  • Weekend-specific content strategies might be beneficial