search

jawadmhoque / transit_ridership_decline

Causes of transit ridership decline
0 stars 2 forks source link

Trend Charts #2

Closed gregerhardt closed 4 years ago

gregerhardt commented 4 years ago

Look at the script outputs/charts. Create the same set for the clusters.

gregerhardt commented 4 years ago

Looking at the start of these trend charts you have in Est6_Outputs, and laying out the tasks, a few things to anticipate:

  1. Within each cluster, the cities with rail are a subset of all cities, but this might only matter for the mid op-ex cluster. For the high op-ex, I think all the cities have rail (?). For the low op-ex cluster, we are excluding the rail. For NY, it is only one city, so both bus and rail. So, what's the best way to deal with this? I think I like how you've done the rail charts where there are two sets of lines. If the set of cities are identical, then the lines should fall perfectly on top of each other for all charts except for VRM and Fare.

  2. It looks like you are plotting the year-over-year change. I think we should plot the actual value instead.

  3. It's worth thinking through whether I've done the aggregation in a sensible way. I think I'm calculating the average value when aggregating to the clusters. Maybe this should be a weighted average or a total? I'm not sure, but is worth thinking about as we go.

  4. It should be the same variables and in the same order as the FAC charts.

gregerhardt commented 4 years ago

What is in the interim report is fine.