I'm using this space as a bit of a shared workspace. So forgive the lack of relevance to the normal postings to this blog.
Click to enlarge.
This is a work in porgress. This particular version is more a proof of concept. Here is what Mr. Walker had to say about my previous version:
Jeff. Thanks! I object, though, to using trips/day as a proxy for frequency, because lines with really intense peak service can have lots of trips/day even with little or no all-day service. . If the point of a frequency map is to show people the network they can travel on without waiting long, it needs to be based on actual midday headway. Cheers, Jarrett
So I asked for some help in coming up with "actual midday headway" thusly:
So if I am to do a midday headway map, I suppose counting the number of trips that leave between 11:30am and 12:30pm would not do because they might send out a bunch of drivers at 11:35 all together and then no more until 12:50 when another bunch might be sent out? So is actual midday headway the time between the first trip nearest noon and the one after that? Forgive my ignorance, this is not an area of expertise, but what is a good mechanism for measuring "midday headway?" What would be a good method of calculation?
To which he replied:
Jeff. You're right, it's tricky!
You have to return to the spirit of Frequent service, and the wholereason that we're mapping it. We want people to see where they can travel ALL DAY without waiting very long.
So the real question is: "What is the longest scheduled gap between consecutive buses anytime during the midday?" The point of frequent service is that it's a guarantee, not an average., That means we are mapping the worst case, not the average case.
While working with the previous version I learned that "Shapes" provided by the CTA were not properties of "Routes" but rather "Trips." In producing this map I further came to realize that "frequencies," or goal here, is also not a property of Routes nor Trips, but rather a property of "Stops".
So using the largest table provided by the CTA, the "Stop Times" table of over two milion separate records, I created a new table that not only provided the time of arrival at each stop but also the time of the previous arrival at that stop. I then calculated the greatest time difference between arrivals between seven AM and seven PM for each stop. I then drew a circle for each stop with larger circles meaning more frequency, or less maximum times between stops. The variance in circle sizes above represent a range of maximum scheduled wait time from seven or eight minutes to thirty or more minutes. Remember that's worse case scenario. I used the same color scheme from the previous map, red and orange buses, green for el/subway and blue for express bus service.
There is a ton of possible tweaking. This is a work in progress. Be patient. I'm trying to be frequent in my updates.