tlumip / swimr

Visualizations and Reports for SWIM
Other
0 stars 1 forks source link

Implement Leaflet #13

Open gregmacfarlane opened 8 years ago

gregmacfarlane commented 8 years ago

It would be good to have leaflet plots of some key variables that the peer review panel can explore. What kinds of things should we visualize? Would love to have input from @bettinardi @rickdonnelly @bstabler.

gregmacfarlane commented 8 years ago

For zones, we could visualize

gregmacfarlane commented 8 years ago

For networks, should visualize

bettinardi commented 8 years ago

I am not fully sure what these would look like, but I would point out that the challenge with the good ideas listed above, is the time dimension. I would also highlight that Ben and I both have produced great tools to visualize zone information (pop, emp, price, floorspace) in the past. The important thing isn't how much there is in a given year in a given scenario. The important thing to review is how these measures are changing over time or across scenarios, so when building these, please make sure the user can do a difference between two years or scenarios. I like the list, one more to consider might be trips and trips by mode, along with economic activity (product) - I bet John would have some great ideas too.

gregmacfarlane commented 8 years ago

I guess we need to have a discussion about what the goal of this r package is, and if it is changing. Originally, it was to create an aid for the team to diagnose issues by having standard plots that we would also show to the peer review panel in the validation testing report. It was not to create a replacement for the interactive and detailed visualizations that @bstabler and @jeabraham have written in the past. I have neither the budget nor the desire to redo their work.

I was moving to leaflet because spatial relationships are important and swimr currently doesn't handle them. I'm staying away from full shiny interaction because at that point, we would have to distribute the VIZ database along with the validation report.

I think @bettinardi's suggestion for elapsed and scenario-diffed images is good and I will do it. I just want to temper expectations about how much I am going to try and swallow.

bstabler commented 8 years ago

Probably the best way to start is to manually create the needed maps in VISUM, and then if we find that we need to recreate them over and over again, automate the process. VISUM's Python API has plenty of features for creating maps, saving jpgs, etc.