Skip to content

Latest commit

 

History

History
60 lines (41 loc) · 5.42 KB

rmed_2020_rundown.md

File metadata and controls

60 lines (41 loc) · 5.42 KB

R/Medicine 2020 Conference - Quick Rundown

R/Medicine was virtual this year, so I finally got to attend. Both R/Med and csv,conf used Crowdcast - my favourite of the many virtual conference platforms I've experienced this year. R/Med was really well organized, with great speakers and a fantastic mix of topics from R, statistics, and clinical studies.

There was a lot of information packed into this 2.5-day conference, so I haven't had time to process everything, but here's a quick list of interesting things that I need to read more about later.

Currently the recordings are available on Crowdcast for registrants, but I've tried to link to open sources below. Check out @r_medicine or #RMedicine2020 for recaps and conversations.

  • Tidymodels/machine learning workshop - this was fantastic, now just need to use this again so I can remember it! We went through an example of building a supervised classification model. Website has links to all the slides and exercises.

  • Reading material:

    • Double dipping in statistical analysis - what is double dipping, and how to fix by conditioning on hypothesis selection procedure. Preprint is still in preparation, but Twitter thread gives a good summary.
    • HealthyR - training & resources for health data analysis
    • Arcus Education at Children's Hospital of Philadelphia - more training & resources

  • Shiny apps:

    • {shinyviz} - template Shiny app for summarizing & visualizing categorical data; customize for your own data
    • {shinyfit} - template Shiny app for linear/logistic/etc. regression analysis; customize for your own data
    • tidyCDISC - Shiny app to explore & visualize ADaM-format clinical trial data; slides here; will be on GitHub "soon"
    • {nDSPA} - Shiny app & package for working with spatial omics data
    • READi Tool - Shiny app for evaluating real-world data - creates custom PubMed search from input; summary here; still in beta testing (app seems to be down as of 9/4/2020)

  • Packages - general R/R Markdown/Shiny:

    • {listdown} - programmatically generate R Markdown files from named lists
    • {dbplyr} - do data manipulation on SQL server, then just pull what you need into R for analysis
    • {drake} - plan your workflow so you don't have to keep repeating the computationally intensive parts
    • {pkgdown} - build a website for your R package (i.e. pkg documentation as knowledge repository)
    • Docker & {holepunch} - share research compendia on GitHub
    • {shinyviz} - template Shiny app for summarizing & visualizing categorical data
    • {shinyfit} - template Shiny app for linear/logistic/etc. regression analysis
    • {golem} - opinionated framework for building robust Shiny apps
    • {pool} to make a separate data layer for your workflow/Shiny app

  • Packages - Publications/communication

    • {gtsummary} - make publication-ready analytical and summary tables. Example:
    • {finalfit} - make nice tables for publication; similar to {gtsummary}

  • Packages - Clinical data

    • {survminer} - make Kaplan-Meier curves with ggplot
    • {ggconsort} - programatically generate CONSORT diagrams; still in development
    • {foreceps} - processing & wrangle ADaM-format clinical trial data; will be available in a few weeks
    • {collaboratoR} - work with REDCap data in R
    • {redcapR} - interact with the REDCap API

  • Packages - Other

    • {treeheatr} - decision tree + heatmap visualization for better interpretability. Example: example from treeheatr github
    • {runcharter} - plot & analyze run charts to track changes over time. Example:
    • {twilio} - connect to Twilio API to send text messages from R