The following are notes from a meeting with @jeffreyhanson, @ricschuster and @edwardsmarc to discuss 2 topics:
What updates would be needed to scale WTD up to work at the landscape scape (i.e. 10,000+ PUs, 4-5 zones/actions)
Are there any other updates we should consider for the next year that would improve WTD for users working at the property scale?
We all agree that resources invested in updates depends on the amount the tool is being used. Currently not many people are using the tool, but usage could ramp up in the next 1-2 years.
1. Updates for landscape scale processing
In order of importance:
1. Map rendering
We would need to improve map rendering since base Leaflet cannot accommodate very large numbers of polygons
Simplest option would be to switch to leafgl for rendering which would give access to fast rendering of WebGL (@DanWismer and @edwardsmarc should test this as our quickest fix).
There may be more work to be done around layer rendering depending on the number of layers we want to display. Currently the TOC is very simple, but we could upgrade to a more sophisticated TOC similar to WTW.
2. Data format
Input data is currently provided in Excel sheets that users manually fill out using a template. With very large numbers of PUs these tables will get many rows which could be a limitation.
Since we'd likely be running these internally in ConsTech we could generate the input sheets in R. Data prep would be internal and users would just be interacting with the app rather than prepping data.
We will likely need to change how the tables are displayed in the app. Moving to paginated tables or hiding tables if they are too large to render effectively.
3. Goals
UI is not designed for large numbers of features.
Goals are currently set manually in the app with one slider per goal. If we start adding larger number of goals we'll need a better system for selecting goals. Something similar to the WTW interface would be one option. Another option would be setting goals in the excel sheets so they're already set when the app opens. Again, for internal data prep this would work fine. We should check how goals are visualized in the app under this scenario.
We'd likely need to condense a large number of features (e.g. species) into a smaller number of layers for use in WTD. e.g. one layer summarising each Theme from WTW.
4. Input data
Challenge at the landscape scale is coming up with meaningful data for actions and outcomes of actions. Dream is to have a database of species specific actions, costs and outcomes based on real world experience that we can feed in to the tool. We're not there yet. Finding proxies for this is a big challenge.
If we can generate pre=prepped input data we could have a standardized data-prep system like we do for WTW.
2. Updates for general usability at the property scale
We generally agree that the tools is useable in its current state for the original intent of property level decision making. The main limitation is getting quality input data (e.g. expert opinions) from users.
1. UI improvements
Sliders can be improved - similar to WTW
The app has an over-reliance on tables for displaying the data. Finding a more effective way to communicate this data would be good.
Depending on user type, we might consider switching the data input formats to something that can more easily be automated. Such as GIS formats. This depends on user preferences, are they GIS users or are they more comfortable filling in the excel tables manually.
The following are notes from a meeting with @jeffreyhanson, @ricschuster and @edwardsmarc to discuss 2 topics:
We all agree that resources invested in updates depends on the amount the tool is being used. Currently not many people are using the tool, but usage could ramp up in the next 1-2 years.
1. Updates for landscape scale processing
In order of importance:
1. Map rendering
2. Data format
3. Goals
4. Input data
2. Updates for general usability at the property scale
We generally agree that the tools is useable in its current state for the original intent of property level decision making. The main limitation is getting quality input data (e.g. expert opinions) from users.
1. UI improvements