Improve Traffic Visualization

[tonywchen]

Problem Currently the map is very sparsely populated and doesn't make it very easy to visualize the traffic

Cause There can be many causes to the problem:

1. The traffic data at any given time are computed on the immediate stops that have an approaching vehicle. This means there can be as many as data points as the number of vehicles service. Even though the current data looks at the average of 5 data points (ie. every 5 minutes), this still doesn't really yield a clear picture of the entire route.
2. The data also do not consider the trickling effect of traffic have on the upstream traffic. The traffic status of a "path" (ie. route between 2 immediate stops) ends abruptly at the starting stop. However, there is an argument to be made that there are too many factors that can affect the upstream traffic, and this might just be the limit of what can be inferred from the transit prediction data
3. The problem can also be on the visualization itself. The visualization shows only a snapshot of the traffic at a time and the resulting visualization can be jumpy and unpredictable and thus results in its ineffectiveness.

Ideas for Enhancements

1. Look further into prediction data and see if additional traffic info can be inferred from the non-immediate stops that have approaching vehicle. For example

   
   Vehicle is approaching A, which is the stop right before stop B. Let's define
   - time_a: number of seconds predicted for vehicle to arrive at A
   - time_b: number of seconds predicted for vehicle to arrive at B
   - duration_ab: number of seconds expected to travel from A to B

Is time_b = time_a + time_b always true, or duration_ab a variable based on some internal insights that Nextbus has.

2. Additional visualization technique to simulate/guess the effect of traffic at a certain path has on it upstream, or the effect of traffic at a certain path has on its own overtime. This would provide a more smooth transition from one data point to the next.