And how…

In college I majored in English and Radio-TV-Film. But somehow I also ended up in three computer programming classes. Those were the ones I enjoyed the most. Among the punch cards I discovered flow.

A decade later I bought an Apple ][. Flow and computers have been a part of my life ever since. I have been in and out of a lot of different computer languages with various degrees of understanding and success, but for the last few years I’ve been learning R, a language designed for statisticians.

R has some advantages and some disadvantages as the technological underpinnings for Open-Meta. To borrow a term from computing, Open-Meta has three threads: to develop and demonstrate a crowd-sourced platform for systematic reviews; to develop and demonstrate a platform for meta-analysis anyone can use; and to seed the platform with a sample systematic review and meta-analysis of randomized controlled trials of daily vitamin D3 supplementation.

R is clearly the best choice for the meta-analysis thread. Others have already written R packages that facilitate meta-analysis. However, these packages assume data collection is complete. R was designed to analyse data, not to collect it.

In computer terms, data collection, the heart of the systematic review thread, is an exercise in CRUD (create, read, update, delete). While there are lots of better languages for writing a program for data collection than R, R is where my head is, so I spent the fall of 2016 writing a program in R for doing the initial part of a systematic review (aka Stage 1 Review). Stage 2, which includes pulling a lot of specific information out of scientific articles and saving that data in a format that R can use, still has to be written.

To add CRUD to R, I used an R package called shiny, which allows you to connect your R programs to input and output fields in a web browser that you customize.

But wait. What you’re reading right now is in a WordPress site on an inexpensive shared server. The shiny programs I’ve written were designed to run only on my computer. I don’t know how complicated it will be, but shinyapps.io provides an inexpensive method of putting my systematic review program online.

My current plan is to display the systematic review pages created on shinyapps.io in an iframe on this WordPress site. But I spend a lot of time wondering whether Amazon Web Services wouldn’t be a better choice.

Leave a Reply

Your email address will not be published. Required fields are marked *