AboutResearchSoftwareTalksTeachingCV

02 / SOFTWARE

Software

R Shiny applicationCreated 2026

Psychiatric Prescribing
Pathways for Bipolar Disorder

Summary Explore 22 years, from 2000 to 2022, of psychiatric prescribing pathways for bipolar disorder in the UK. Sidebar controls compare initial treatment regimens, focus on lithium, change concurrency windows, and export selected visualisations. The interactive Sankey diagram provides detailed information on the percentage of patients switching treatments.

Implementation Developed using R Shiny.

Launch Shiny app Read the research
Open-source web repositoryCreated 2026

UCL DoP Phenotype
Codelists Repository

Summary In UK electronic health records, researchers must define codelists that specify which medical codes correspond to each health condition. With thousands of non-standardised and frequently updated codes, creating and maintaining codelists is time-consuming and error-prone.

This repository provides an open-source catalogue of phenotype codelists, reproducible code for generating them, and metadata documenting their construction and clinician validation.

Implementation Python automatically pulls relevant codelists from contributing GitHub repositories created by clinicians and researchers in the UCL Division of Psychiatry Mental Health Data Science Team.

View live platform Source on GitHub
Created 2024
Pattern probabilities plot generated by the baysc R package

R package

baysc: BAYesian Survey Clustering

Summary An R package to run Bayesian supervised and unsupervised clustering methods to find underlying latent classes in survey data while accounting for complex survey design. The package vignette provides a step-by-step guide to installation, model fitting, plotting and summarisation, and performance diagnostics.

Implementation Developed using R, Stan, and C++.

View repository
Created 2021

WHO · R Shiny applications

WHO HIVDR Sample Size Calculators

Summary Calculate required sample sizes for national-level surveys of HIV drug resistance prevalence around the world. Laboratory method I supports countries with viral-load coverage of at least 60% and required survey variables of at least 80%; laboratory method II supports coverage of at least 60% and variables below 80%; the clinic-based method supports coverage below 60% and variables below 80%.

Implementation Developed using R Shiny.

Laboratory method I Laboratory method II Clinic-based method

LEARNING RESOURCE · CREATED 2025

Building Your Own R Package

A practical tutorial developed for the UCL R User Group on structuring, documenting and sharing an R package.

Open tutorial

LANGUAGES & TOOLS

01R
02Python
03SQL
04Stata
05MATLAB
06SAS
07Java
08GitHub
09LaTeX