I’ve had R Shiny app development on my list of things to do for a loooong time. I think I’ve managed to wait long enough to ensure that even 4th graders now have better data science resumes than me (which probably means that there is something newer and way cooler waiting just barely outside the public conciousness) but I’m not going to let that stop me from getting this crossed off my list.

Shiny is a reall cool module for R that allows you to build applications and deploy them in a variety of formats. I’m interested in these for two reasons:

  1. Decision support - I work on a lot of projects where I am part of a research team tasked with delivering science advice to policy makers (presumably because they want to make smart and informed public policy decisions that reflect the best available understanding of current information). Policy makers and managers often do not want to see the guts of a carefully crafted econometric/statistical/applied mathematical model. However, scientists often do not want to provide managers with the kind of ‘do this/don’t do that’ recommendations they crave because the real world is complex, even the best analysis has some uncertainty associated with it (fundamental weaknesses in the data that arise because of the messiness of collecting data in the real world, analytical uncertainty that comes from modeling noisy data with statistical randomness, etc.)

  2. Democratization of data - I work on a lot of projects destined for the peer-reviewed scholarly scientific literature where, if I’m lucky and have an idea with a massive impact and huge following, it might get read by 100 people. I like the idea of making