Monkey Pox and Shiny Maps
Health Data Weekly — Week 34
The rapid spread of Mpox accross Africa is alarming. With 18,910 cases and 541 deaths reported in 12 AU Member States — this calls for urgent action and mobilization of resources.
— Cyril Ramaphosa, President of South Africa.
Editorial
‘Infectious Disease’ is probably not the first thing people want to read about nowadays. The spread of mpox -also known as Monkeypox- has become a health crisis in the Democratic Republic of the Congo. With reports that the disease has made its way to South Africa and recently Sweden there is a lot of panic going on right now.
While the WHO have declared mpox a Public Health Emergency of International Concern, many social media commentators are shrugging off the mpox reports. Talking points include claims that ‘mpox is only transmitted through close contact,’ or that ‘mpox has a low fatality rate’. And worse yet are the accusations that ‘mpox disease is fake and the pharmaceutical industry is diseasemongering’.
Talking through one’s hat, it’s easy to say anything and everything about mpox. But we can do much better. The WHO have created a Shiny page for following mpox as data as we are learning more about it: https://worldhealthorg.shinyapps.io/mpx_global/
This above figure on mpox cases over time also comes from said mpx_global Shiny page. It paints the very interesting picture that the US, Brazil and Spain have quashed their respective epidemics already, months ago. That explains all the staunch confidence we keep seeing on social media. On the other hand what is happening at this time is mpox spread picking up momentum through wholly different contact points in Africa.
As far as Africa is concerned, mpox is to them a growing continental epidemic. And it matters.
JYNNEOS debuted last year as Bavarian Nordic’s US CDC-approved mpox vaccine. And now, African countries are struggling to obtain Jynneos because the US & Europe hoarded jabs. On the bright side, mpox veterans Spain and the US have made pledges to donate from their national stockpiles.
I would like to think that once again, these Shiny Maps are an example of why we need data participants instead of data subjects. It’s very easy to look at a bar chart and say ‘my country is doing fine enough’ and move on with our day. But that’s not how diseases work, because they do not recognise the same borders. If African voices were promoted sooner, these countries would have been allowed to prepare in the same way that the US and Europe have. We can only achieve data participation through humanising one another, acknowledging experiences and delivering treatments not where they are most wanted but where they are most needed.
AI
Bayesian Online Neural Networks with Learning Machine: In modern AI we have two problems, Variability of Results and Model Updates. Thierry Moudiki’s latest update to the Learning Machine package will introduce a Bayesian Predictor: bayesianrvfl. That’s an abbreviation for Bayesian Quasi-Random Functional Link.
Bayesianrvfl both predicts a distribution (rather than a single point) and is fairly easy to update with successive samples in the future (instead of traditional batch-training and batch-prediction).
Broader Data Science
A New R Community in Ahmedabad, India, focused on Clinical Research and Pharmaceutical Industries: Ahmedabad is now emerging as a significant hub for clinical research, data sciences, and pharmaceutical innovation. Read on to learn more!
Guides
Plot Your Life in Weeks: The original Your Life in Weeks, by Tim Urban was published in 2014.
Using R to prompt a local LLM with ollamar: Tutorial on using ollamar for learning about genes or summarising the latest genetics papers on bioRxiv.
Unveiling Bottlenecks: A Guide to Profiling R and R Shiny Code: Profiling tools can help reveal what functions/processes are draining on resources.
Quarto dashboard creation and automation: Quarto is a great format for writing code, documentation that you later want to publish. Quarto now supports dashboards — a new output format for easily creating dashboards from notebooks in R, Python, or Julia.
R tidymodels: A tidyverse Like Ecosystem for Efficient Machine Learning in R: A nice tutorial on tidymodels, one the state-of-the-art ML frameworks for R.
Events
Navigating the R ecosystem using R-universe: Virtual. Tuesday, 24 September 2024 16:00 UTC. Virtual Session on learning how to use R Universe.