Michael Levy

@taylorlarsen -- I think the `evaluate` package will do what they need. It will still require some parsing of the output, but it looks pretty straightforward. Check out the `evaluate_demo.R`...

Currently hyperparameters trump, e.g. ``` flash_models(dd, diabetes, models = "rf", hyperparameters = get_hyperparameter_defaults("knn", nrow(dd), ncol(dd), "classification")) ```

Yes to patchwork. Thomas says he's planning to put it on CRAN after ggplot 2.3 release.

@mmastand Putting this on your radar since you're working with this. Is there a use case for us connecting to MSSQL from Mac?

Can we setup Zenodo to auto-pull new releases? It's currently back at v1.0.

Thanks for thinking through this and documenting it well @levithatcher. Connected to #857. This is @mmastand's awesome function, so he should weigh in. Thinking about how we handle them in...

@aadejare The package currently doesn't support multiclass prediction, but we plan to add this functionality in the next month.

```r pima_diabetes % dplyr::mutate(diabetes = dplyr::case_when( diabetes == "Y" ~ TRUE, diabetes == "N" ~ FALSE)) machine_learn(pima_diabetes, patient_id, outcome = diabetes) ```

@MaxTaggart I wouldn't install R with homebrew. Would you mind installing via the installer and let me know how that goes: https://cran.r-project.org/bin/macosx/

We definitely need functionality like this. I've left `selectData` and `writeData` in for now. I don't know how db connections work well enough to rewrite them well.