Temporal Patterns Analysis

Understanding how Citi Bike usage varies by time of day, day of week, and season around Columbia University.

Research Question

How does Citi Bike usage vary near Columbia University by season, weekday, and time of day?

Overall Statistics

Total Trips

529,093

From 2024-01-01 to 2025-10-31

Average 789.7 trips/day

User Distribution

Members

82.4%

435,971 trips

Casual

17.6%

93,122 trips

Trip Duration

Median:9.4 min
25th percentile:5.2 min
75th percentile:17.6 min

Trip Characteristics

Understanding trip duration patterns and the distribution of user types and bike types provides insights into how the system is being used.

User Type Distribution

Loading chart...

82.4% members vs 17.6% casual riders. The high member percentage suggests the system is primarily used for regular commuting rather than tourism or occasional use.

Bike Type Distribution

Loading chart...

79.8% electric bikes vs 20.2% classic bikes. The strong preference for electric bikes reflects Columbia's hilly terrain and the convenience of e-assist for longer or more strenuous rides.

Trip Duration Distribution

Loading chart...

Insight: The median trip duration is 9.4 minutes, with most trips falling between 5.2 and 17.6 minutes. This log-normal distribution is typical of bike-share systems, where most trips are short commutes or errands, with a long tail of longer recreational rides.

Hourly Usage Patterns

Understanding when bikes are used throughout the day reveals clear commute patterns with peak usage during morning and evening rush hours.

Loading chart...

Key Insight

Peak hour is 17:00 with 47,217 trips, indicating strong commute patterns aligned with typical work/school schedules.

Activity Heatmap: Day × Hour

This heatmap reveals usage patterns across different days of the week and hours of the day, showing distinct weekday commute patterns versus weekend recreational usage.

Loading chart...

Key Observations

  • Weekdays show clear morning (7-9 AM) and evening (5-7 PM) peaks
  • Weekend usage is more spread throughout midday hours
  • Night hours (10 PM - 6 AM) show consistently low usage across all days
  • Wednesday and Thursday typically show highest weekday activity

Member vs Casual User Patterns

Comparing hourly usage between member and casual riders reveals distinct behavioral patterns. Members show strong commute peaks, while casual users exhibit more leisure-oriented usage.

Loading chart...

Key Observations

  • Members show sharp peaks at 8-9 AM and 5-6 PM (typical commute times)
  • Casual users have more consistent usage throughout the day
  • Casual usage peaks in afternoon/evening (12 PM - 6 PM)
  • Member usage significantly exceeds casual usage during rush hours

Usage by Time Period

Categorizing trips into time periods (Morning Rush, Midday, Evening Rush, Night) shows the distribution of usage throughout the day.

Loading chart...

Key Insight

Evening Rush (4-8 PM) shows the highest usage, followed by Midday and Morning Rush. Night hours (8 PM - 6 AM) account for the smallest share, consistent with typical urban bike-share usage patterns.

Weekly Patterns

Trip volumes vary across days of the week, with clear differences between weekdays and weekends reflecting Columbia's academic calendar and commute patterns.

Loading chart...

Key Observations

  • Weekdays (Monday-Friday) show consistently higher usage than weekends
  • Mid-week days (Tuesday-Thursday) typically have highest volumes
  • Saturday shows lowest usage, followed by Sunday
  • Pattern aligns with Columbia University academic schedule

Weekday vs Weekend User Behavior

Breaking down usage by day type and user category reveals how different user groups contribute to weekday and weekend ridership.

Loading chart...

Key Insight

Members dominate both weekday and weekend usage, but the gap narrows on weekends. Casual users represent a larger proportion of weekend trips, suggesting more recreational usage patterns compared to member commuting behavior.

Monthly Trends

Tracking ridership month-by-month from January 2024 to October 2025 reveals seasonal patterns, growth trends, and the impact of weather and academic calendar.

Loading chart...

Key Observations

  • Strong seasonal variation with summer peaks and winter lows
  • Peak months show 2-3× the ridership of lowest months
  • Dips in December/January correlate with winter break
  • May-October consistently show higher usage (favorable weather + academic year)

Seasonal Patterns

Aggregating data by season highlights the significant impact of weather conditions and academic calendar on Citi Bike usage around Columbia University.

Loading chart...

Key Insight

Summer shows the highest total trips, followed by Fall. Spring shows moderate usage, while Winter has the lowest ridership. This pattern reflects both weather conditions (temperature, precipitation) and Columbia's academic calendar (summer courses, breaks).

Hourly Patterns by Season

Examining how hourly usage patterns shift across seasons reveals the interaction between time of day and seasonal factors.

Loading chart...

Key Observations

  • Peak hours remain consistent across seasons (morning and evening rush)
  • Summer shows higher usage across all hours compared to other seasons
  • Winter shows the most dramatic peaks/valleys (concentrated commute usage)
  • Fall and Spring show similar patterns, bridging Summer and Winter extremes

Key Findings Summary

Time of Day Patterns

Usage shows clear commute patterns with peaks during morning (7-9 AM) and evening (5-7 PM) rush hours. Member users drive weekday commute peaks, while casual users show more midday and weekend activity.

Day of Week Patterns

Weekday usage dominates, consistent with Columbia University academic and commute patterns. Weekend usage is lower but shows different hourly patterns with later starts and more spread throughout the day.

Implications for Infrastructure

  • Peak demand periods require sufficient bike and dock capacity
  • Rebalancing operations should focus on morning and evening commute times
  • Member vs casual patterns may require different station configurations
  • Consider dynamic capacity adjustments based on time of day and day of week

Methodology Note

Trip duration outliers (< 1 minute or > 180 minutes) were removed to ensure data quality. Analysis uses median values for trip duration due to log-normal distribution. All visualizations are interactive - hover for details, zoom and pan to explore.

View Detailed Methodology →