Tom Liptrot

About | Apps | Publications | Talks

About me

I am a Data Scientist working at the Christie Hospital in Manchester. The Christie is one of Europe's largest specialist cancer hospitals. My work in involves finding ways to use large amounts of data to improve patient care. I provide statistical modelling support to clinicians conducting clinical cancer research.

My interests lie in the analysis of large complex datasets and I am conducting research into text-mining of electronic patient records collaborating with Manchester University. Techniques I use include elastic net regularised regression, Bayesian hierarchical modelling, GAMs, GLMs and parametric and non-parametric survival modelling. I use the R programming language for most of my analysis.

I have made some shiny apps, I have published some papers and I have given some talks.

I am on Github and Linkedin


Shiny Applications

Shiny is a web application framework for R. It allows users to quickly create interactive web applications based on analysis done in R. I have created a couple of publicly available shiny apps and an R package that makes it even more simple to create predictive shiny apps.

Drug reponse

IL2 app

This app is for predicting patient response to the kidney cancer drug IL-2. It is based on data from 145 metastatic renal cell carcinoma patients.

Job Search

International Jobs app

This app is for finding a job anywhere in the world. It connects to the Indeed API and searches all 54 international indeed websites, returning the results on a map of the world

R Package


Predictshine is an R package that allows users to create interactive shiny based web-apps to make predictions about individuals. its main function predictshine() works similarly to the predict() function.



A selection of some of the publications I have collaborated on:

Evaluation of an automatic segmentation algorithm for definition of head and neck organs at risk

— David Thomson, Chris Boylan, Tom Liptrot, Adam Aitkenhead, Lip Lee, Beng Yap, Andrew Sykes, Carl Rowbottom and Nicholas Slevin. Radiation Oncology 9.1 (2014): 1. link

Patient Characteristics are not Associated with Clinically Important Differential Response to Dapagliflozin: a Staged Analysis of Phase 3 Data

— Sarah Bujac, Angelo Del Parigi, Jennifer Sugg, Susan Grandy, Tom Liptrot, Martin Karpefors, Chris Chamberlain, Anne-Marie Boothman. Diabetes Therapy 5.2 (2014): 471-482. link

Retail gasoline pricing: A Bayesian hierarchical approach to modeling the effect of brand on elasticity

— David McCaffrey, Tom Liptrot, and Barbara Jenkins. Journal of Revenue & Pricing Management 10.6 (2011): 514-527. link

Optimization of forecourt fuel pricing

— David McCaffrey, Tom Liptrot, and Barbara Jenkins. Journal of IMA Journal of Management Mathematics (2011): dpr017. link

My Google Scholar page



Machine learning can predict neutropenic sepsis in chemotherapy patients

— A presentation about a machine-learning algorithm that uses routinely collected clinical data to accurately predict which chemotherapy patients will get neutropenic sepsis. Delivered at Cancer Data and Outcomes Conference. June 2016.

pdf of slides

Abstract in conference programme

Machine Learning in R: Supervised Classifciation

— A talk about some machine learning algorithims at the Manchester R user group. Feb 2016.

pdf of slides

code on Github

predictshine or "How I learned to stop worrying and make an R package"

— presentation at the Manchester R user group. June 2015.

pdf of slides

What will happen to this patient? Predictive text mining in a specialist cancer hospital

— presentation at the EARL Conference, London. September 2015.

pdf of slides