I am currently working to develop an interactive interface of an R machine learning model we're developing. What we're doing is pulling data down from Domo, conducting machine learning, and then pushing two outputs:
- Predicted values for existing data back to Domo (NOTE: This works seemlessly)
- Predicted values for dynamically generated user inputs (NOTE: This is the struggle point.)
Currently, I'm working with Shiny to make the interface, which is an R package. I was wondering if there was similar functionality with Domo as that is what our users will be used to using.
In case the description isn't clear:
- Imagine you have a two data sets in Domo
- An existing data set of of independent variables and dependent variable(s)
- An "new" data set of just independent variables
- We are pulling this data set from Domo into R, where we are conducting machine learning to predict the dependent variable(s) for the new data set.
- We then push the predicted values for a new data set into Domo. (NOTE: All of this works so far.)
- Finally, we are building an interface in Shiny (currently) where users can input their own data and dynamically see the results for that one prediction. This is the part that I don't know how to do. Is there a way to make Domo the interface instead of Shiny?
The crux of the issue is that we will already have security groups/access levels/etc. in Domo, which is to say nothing of interactions between users.
