# Various Analyses of the Data

## Boston Highway Travel Times, July 2002 - May 2004

I have done a further analysis on the data, from July 2002 to May 2004. Using the graphs of each route by month, I have examined the peak of the morning and evening rush hour, the travel time at the peak of rush hour, and the duration of rush hour. Here are the routes I have analyzed:
• I-93: Route 932. "Rush hour" is when the travel time is >= 20 min.
• I-90: Route 90. "Rush hour" is when the travel time is >= 55 min.
• Rte 128: Route 953. "Rush hour" is when the travel time is >= 30min.
I only analyzed the rush hour direction; that is, inbound to the city in the morning and outbound from the city in the evening.

The "rush hour" limit is roughly 10 min. more than the speed-limit travel time. In the table, I am listing the rush hour beginning and end, assuming that the duration of rush hour is symmetric around the peak time. In actuality, the peak is usually shifted toward the end of rush hour.

RoutePeak TimePeak Trip Time (min)Duration (min)Start TimeEnd Time Peak TimePeak Trip Time (min)Duration (min)Start TimeEnd Time
9328:37am292056:54am10:20am 5:21pm271993:41pm7:01pm
908:38am671897:03am10:13am 5:51pm701754:23pm7:19pm
9538:24am32577:55am8:53am 5:37pm381494:22pm6:52pm

Plots of this information by month are shown below. Month #1 is July 2002. The left column is for morning rush hour and the right is for evening rush hour. In all plots, the white represents I-93, the red I-90, and the green Rte 128. The first row of plots shows when the peak of rush hour occurs: 540 minutes is 9am, and 1020 minutes is 5pm. The second row of plots shows what the longest travel time is during that rush hour. The third row of plots shows the duration of rush hour in minutes.

Some interesting observations can be made:

1. Morning rush hour on I-93 and I-90 are very similar. Their rush hours are quite long.
2. Evening rush hour on I-90 and Rte 128 are similar. Their rush hours start somewhat late.
3. Evening rush hour on I-93 starts early and is very long.
4. Morning rush hour on Rte 128 is surprisingly short and begins relatively late.
5. For I-90, the when the worst of rush hour occurs during the morning seems to oscillate between earlier and later times. I do not know if this is statistically significant. The same behavior is seen for I-93 for the evening rush hour.
6. The behavior of the plots of the travel time for I-93 and Rte 128 are very similar for both rush hours.
7. The travel times for I-93 and Rte 128 are fairly constant over these two years, but the travel times for I-90 seem to be gradually increasing, especially for the evening rush hour.
8. The duration of the evening rush hour for I-90 seems to be increasing.

## Plots of Average Travel Times, July 2002 - May 2004

I have plotted the average travel times for the AM rush hour inbound and outbound travel times and the PM rush hour inbound and outbound travel times. (Remember that above I plotted the peak, or worst-case travel times.) The data included were from July 2002 - May 2004. 1-sigma errorbars in the times are included on all points.

Some of my comments and conclusions:

• Many of the curves show a sharp drop in travel times at the beginning of 2003 which continues until about Sep 2003. Was this due to people being laid off, or did everyone take a semester off from school?
• Over all, the travel times seem remarkably constant, despite Big Dig efforts. Some trends in the data:
• The I-93 inbound PM rush hour may be getting shorter.
• The I-93 inbound AM rush hour has a big increase in April 2004 -- I don't know why.
• The Rte 128 outbound PM rush hour seems to be getting shorter.
• While the Big Dig has opened both new tunnels in the Central Artery, I'm not convinced they have helped I-93 or I-90. The inbound AM rush hour travel times are not substantially improved.
• I-95 has a sharp increase in inbound AM rush hour travel times in both Sep 2003 and Sep 2004. Apparently we are all eager to get back to school.
• The I-95 and Rte 24 inbound AM rush hour curves mirror the I-93 inbound AM rush hour curve. Since many commuters use Rte 24 especially to get to I-93, it shouldn't be surprising that I-93's delays propagate back to Rte 24 and I-95.

Travel Times main page
My homepage
avondale@astro.umd.edu