Loading tm in 082-word-cloud causes NLP::annotate to mask ggplot2::annotate

Which causes problems when flipping through the example apps in the same session, for example

```r
shiny::runApp("057-plot-plus-three-columns")
shiny::runApp("082-word-cloud/")
shiny::runApp("139-plot-brush-scaling/")
```

Probably the best way to address is to unload tm/NLP after this app's session ends