Month of Year Delay Pattern

• 2012-2016 are alike across months and types of delay causes.
• Delay minutes increase from May to August, and decrease from August to November. December does not follow this trend, demonstrating instead longer delays than in November. Delay minutes are relatively stable through January to April.
• Delay minutes are longer in summer (peak in August) than in winter. This pattern is possibly due to the fact that hotter air gets thin, making it harder to take off and land safely, mostly for smaller jets. (www.phys.org)
• The graph shows a slightly difference on late aircraft delay between years. A t-test is performed to test if the difference is significant between the beginning and the end of this 5 years range. (see following T-test)

T-test

• The t-test suggests a non-significant difference on late aircraft delay between the 2012 and 2016 at 5% significance level (95% confidence interval, -3.22 to 1.83; P = 0.3588).

Day of Month Delay Pattern

Each trendline represents a month of the year, and each data point on the corresponding trendline represents the average of 5 years’ of delay data (in minutes) stratified by day of the month. (30 days, 30 data points per month).
This graph is created to highlight possible association between holidays and delay minutes.

• August 8th and 9th have the respective longest and second longest delay time of the entire year.
  
• December 17th has the longest delay in December.
• November 17th and 21th have equally longest delay.
• It can be speculated that the December peak is due to the winter holiday season, and the November peak is likley due to the Thanksgiving holidays.

Weekly Delay Pattern: Delay Minutes of Different Causes

Each bar represents the average of 5 years’ of delay data stratified by day of the week. Each bar for day of the week can be broken down into various causes of delay and each corresponding amount of delay due to that reason.

• No difference in total delay minutes across the week.
• No big difference in each cause of delay across the week, except for security delay.
• From a relative perspective, security delay on weekends is twice as long than on weekdays. However the absolute difference is pretty minor, only around 0.5 min.

24H Delay Pattern: Proportion of Delay & Delay Time in Minute

The graph shows what proportion of each hour of flight departure time is delayed. 5 years’ of delay data is stratified by hour of the day. Each data point represents the time interval of 00-59 minutes of each hour, condensed into one data point per hour over 24 hours. Bubble size indicates delay minutes. The longer the delay time per delayed flights, the larger the bubbles are.

• From 00:00 to 03:59, the delay proportion are highest, around 0.5, and these delays are also the longest of the entire day (00:00-24:00), around 150min. Better not to take flights depart at this time range.
• The delay proportion decreases dramatically from 04:00. However, the delay minutes are relatively long from 04:00 to 04:59 compared to the following hours.
• From 05:00 to 23:59, the delay proportion gradually increases and delay time gradually increases as well.