@pgdataviz.bsky.social
24 followers 24 following 25 posts
Statistical Programmer (Epidemiologist by training) using #Rstats and #Python mostly. #DataScience #Dataviz https://pgdatavizandstats.netlify.app/
Posts Media Videos Starter Packs
Holiday and job-hunting sucking my soul mean I am a week behind and needed to do a(nother) rush job for #TidyTuesday. I have been playing with Reactable though

Code here: pgdatavizandstats.netlify.app/data_visuali...

#rstats | #dataviz
Thank you. I would love to be able to "render" the pdf like you can in RStudio.

I'll have a look. Thanks again
I do like #positron but RStudio allows me to create .tex files from new.

Anyone know how I could get this on #positron, or is not a thing yet?
Tried the datapasta package out, which is a form of minor miracle to get the country codes for the players.

Managed to crack a map of grandmasters per country before flying home to the UK for a break

Code here:

pgdatavizandstats.netlify.app/data_visuali...

#dataviz | #rstats | #tidytuesday
A LLM trained on UK Biobank data to predict future disease presented uncritically by the media?
Missed last week's #TidyTuesday / #pydyTuesday due to work/toddler commitments but I'm ahead for this one. Such is life

Tried making a dashboard in python Whoever coded #streamlit is a genius

Dashboard here

tidytuespassportdashboard.streamlit.app

Code here

pgdatavizandstats.netlify.app/datavis
If the 18 month old daughter allows me to, I'm going to take a look this weekend
For #TidyTuesday / #PydyTuesday I have struggled with the port of patchwork over to python. If anyone has any ideas on how to fix the titles of the graphs, I'm all ears.

Also Plotnine is a minor miracle.

Code Here:
pgdatavizandstats.netlify.app/data_visuali...

#dataviz | #python
I logged on, opened RStudio, and accidentally pressed CTRL+SHIFT+P, and what sort of magic is this?
(Reposted because I can't spell) Back to the warm blanket that is ggplot2 for #TidyTuesday.

A Sankey diagram of how the classifications of Scottish Munros have changed over time

#rstats | #dataviz | #ggplot2

Code here:

pgdatavizandstats.netlify.app/data_visuali...
In a vain attempt to keep learning python I have tried imputing some CDISC data using #Quarto and so on - the report is rough around the edges but at least it worked.

#rstats | #python | #imputation

Code Here: pgdatavizandstats.netlify.app/biostatistic...
Another #TidyTuesday / #PyDytuesday another week of thinking that making graphs in Python just isn't as nice as it is in R (Skill issue, I know)

Code here:
pgdatavizandstats.netlify.app/data_visuali...

#dataviz | #python
I've been cursing out my Propensity Score Matching project all week at work so a very brief #TldyTuesday and #PyDyTuesday. If I have time I'll try and do a better job.

Code here: pgdatavizandstats.netlify.app/data_visuali...

#python | #dataviz | #rstats
Anyone got any useful guides on Network Meta-Analysis in R?

Context I have done 1 in my life and despised it and swore I would never do one again. Alas that was not to be.

If I could avoid WINBUGS entirely that would be great

#rstats #episky
Created an Interactive map with Folium of all the New York MTA artworks for #TIdyTuesday and #PyDyTuesday

#python | #dataviz | #gis

Code here: pgdatavizandstats.netlify.app/data_visuali...
It's not perfect and I can't switch back and forth from landscape to portrait but it does mean a much nicer report
One of things that's stopped me using Quarto full time as opposed to officeR and an R script is the need to use loops to repeat image and table production.

I've now worked out how to run a loop within an R code chunk and images, tables as the loop runs so I get a nice doc

#Quarto #rstats #python
Back to ggplot2 because I was in a rush. British Library Funding

Significant downturn in funding post 2010. As prospective employers may see this one day, I don't speculate on the reason.

code here: pgdatavizandstats.netlify.app/data_visuali...

#RStats | #TidyTuesday | #DataViz
So if I build the quarto doc in studio it will render fine.

If I do it in positron I will get issues like "can't find quarto version on your pc".

Mate it's there it works on RStudio. Fiddly stuff like that that I could probably solve but I just find it easier to fire up RStudio
I do use it and would like to use it more, but I use Quarto/RMarkdown a lot and RStudio runs better with those two. I tend to get errors that I wouldn't rendering the same doc in RStudio.
Anyway, tried something a bit more advanced this week, building a Model in Python to predict petrol (Americans it's a liquid - your name is stupid) prices.

Borrowed heavily from here shorturl.at/3q1BR

code here: pgdatavizandstats.netlify.app/data_visuali...

#Python | #TidyTuesday | #PyDayTuesday
So I've split out the "Data Science/R Programmer/Python" part of my "main" profile because I found the fact people may be following me for Data content and just seeing relentless posts about Welsh rugby.