Michał Wypych
LightNews LightNews
mwypych.bsky.social
Michał Wypych
@mwypych.bsky.social
150 followers 1.1K following 72 posts
R / Shiny dev
Background in academia
#rstats | #python | #dataviz | #rshiny

https://mic-wypych.github.io/
Posts Replies Media Videos
mwypych.bsky.social
Basketball #tidytuesday

Adventures in {grid} part 2, tried to make a basketball court.

code: github.com/mic-wypych/t...

#rstats | #dataviz | #ggplot2
A plot showing eurobasket teams that made it to final four or won a title on plot styled as a basketbal lcourt
October 9, 2025 at 8:19 PM
mwypych.bsky.social
only a week late to #Tidytuesday!

I started learning {grid} and while it's great its. so. many. iterations. One thing I learned so far: always draw by hand before coding!

#rstats | #dataviz | #ggplot2
A combination of heatmap and stacked bar chart showing share of men & women chess players across games played and ranking in september 2025. Data from FIDE
October 3, 2025 at 8:40 AM
mwypych.bsky.social
#tidytuesday

I looked at most active frog spotters and made a hopping cumulative plot.

code: github.com/mic-wypych/t...

#rstats #dataviz #ggplot2
September 3, 2025 at 8:42 PM
mwypych.bsky.social
Another #tidytuesday.

I looked at inequality data for Poland from 1986 to 2023.
This one was largely inspired by W.E.B. Du Bois.

code: github.com/mic-wypych/t...

I love {ggforce} and the myriad options it opens.

#rstats #dataviz #ggplot2
A plot showing gini coefficient in Poland from 1986 to 2023 and major events from Polish history from that period. Gini (tax adjusted) was generally around .3. The plot is red and black, in style inspired by W.E.B. Du Bois
August 8, 2025 at 7:38 PM
mwypych.bsky.social
Yes! Mine looks like this:
August 7, 2025 at 5:57 AM
mwypych.bsky.social
#tidytuesday

Did shows on Netflix get longer with time? Not really but interestingly they cap at 60 hours with very few longer shows.

code: github.com/mic-wypych/t...

#rstats #ggplot2 #dataviz
a plot that shows runtime in hours vs release date of netflix shows. Just a few are longer than 60 hours but a lot are just under 60 hours with majority being much shorter (under 25 hours)
August 3, 2025 at 7:46 PM
mwypych.bsky.social
#Tidytuesday

I looked at categories of funding of British Library. Sadly, all of them are down in 2023 with the peak in funding in 2006.

code: github.com/mic-wypych/t...

#rstats | #ggplot2 | #dataviz
A plot showing changes in funding of British library through categories of funding: GIA, investment, voluntary, services and others from 1998 to 2023. Sadly funding across all categories is down.
July 19, 2025 at 8:13 PM
mwypych.bsky.social
#tidytuesday about colors from the xkcd color survey!

Colors are always fun to work with. I looked at overlaps in descriptions of colors given by participants of the survey.

code: github.com/mic-wypych/t...

#rstats #dataviz #ggplot2
a plot showing overlaps in descriptions of colors (purple, pink, blue, brown and green) from the xkcd color survey
July 9, 2025 at 7:59 PM
mwypych.bsky.social
This #Tidytuesday I looked at weekly gasoline prices in EU and US and the 2014 and 2022 invasions by Russia. Data for EU grabbed from EU weekly oil bulletin

If you want to support Ukraine and brush up some data skills consider Workshops for Ukraine by @dariia.bsky.social!

#ggplot2 #dataviz #rstats
The plot shows a trend of weekly gasoline prices in EU and US from 2013 to 2025 and highlights Russian annexation of Crimea and Russian invasion from 2022. While there was little impact in 2014 there was a significant spike in prices after the 2022 invasion started.
July 1, 2025 at 8:55 AM
mwypych.bsky.social
For this #TidyTuesday I made an animation showing measles lab confirmed cases per 100k in EU.

Animation with {gganimate}, legend added with weird {magick} tricks.
I'm not fully satisfied with this one, working with animations is tough!

code: github.com/mic-wypych/t...

#rstats #ggplot2 #dataviz
June 25, 2025 at 7:44 PM
mwypych.bsky.social
#tidytuesday week 24

I was curious about most common licenses used in the APIs, their categories and providers. Made simple plots but wanted to combine them together.

code: github.com/mic-wypych/t...

#rstats #dataviz #ggplot2
The plot shows MIT, Apache, Creative Commons, Other and missing licences from APIs guru data for Tidytuesday 2025-06-17. The plot is divided into two plots and a table. The left plot shows when of APIS in each category were added to APIs guru. CC and Apache licenses were added mainly post 2020. A lot of APIs with missing license were added before 2020. The right plot shows most common categories of APIS in each license type. The table shows most common providers for each type of API license
June 17, 2025 at 10:16 AM
mwypych.bsky.social
#tidytuesday week 23.

I looked at judge appointments by US presidents and parties.
The record goes to Reagan with 387 nominations (both terms).

Data with beginning and end of term for each president are from Kaggle US presidents dataset.

code: github.com/mic-wypych/t...

#rstats #ggplot2 #dataviz
the plot shows judge appointments by US presidents from Washington to Obama. Color codes the party the president is from. Reagan appointed the most judges (387), Clinton was second (383). There were much fewer judges appointed in the XIX Century than in XX Century.
June 11, 2025 at 7:41 PM
mwypych.bsky.social
Thanks for pointing it out! I messed up when cleaning the code and removed a condition that deals with situations when n in comb() is lower than 2. I updated the code on github and below is a screenshot of the function. Lines 6-8 is where the difference is.
June 6, 2025 at 8:53 AM
mwypych.bsky.social
#tidytuesday, this time about the gutenberg project!
I looked at coauthors - what are the age differences and what are the most common collections of coauthored books.

For me this one was a lot more about wrangling the data than making a plot.

code: github.com/mic-wypych/t...

#dataviz #rstats
a histogram showing age differences of coauthors from gutenberg project. The average difference is 29.6 years but the most common difference is 5 years. The plot also shows a treemap with most common collections of coauthored books. Most common ones are culture, literature, young adult and history.
June 5, 2025 at 12:16 PM
mwypych.bsky.social
This #Tidytuesday is about the 2025 #UseR conference!
I made a simple interactive plotly schedule for the event. You can browse the talks and look up the dates, times, speakers and abstracts.

Working version: mic-wypych.github.io/user2025_sch...

code: github.com/mic-wypych/t...

#rstats #dataviz
An interactive plotly that shows talks in each room by time slot and date at 2025 useR! conference. The plot allows to hover above each talk to get more information like the speakers, title, date and time, description and information whether video recording will be available. The plot is made in cold blue shades with black text
April 30, 2025 at 8:58 AM
mwypych.bsky.social
Made my first #quarto extension with #Typst!

Does anyone know how to store a quarto extension (and use it in a team) if I can't put it as a public repo on github? Ideally it would still be in version control (e.g. can I keep it on private github repo and push as zip to some google drive?)
#rstats
a screenshot of frist 6 pages of a polish report made with quarto and Typst
April 16, 2025 at 8:01 AM
mwypych.bsky.social
Little late to #tidytuesday. This time I made a simple ggiraph that shows ranks of each state in various waiting time measures.

code: github.com/mic-wypych/t...

#rstats #dataviz
an interactive map showing ranks of all states in various medicaid waiting time measures for tidytuesday 2025-04-08
April 12, 2025 at 9:40 AM
mwypych.bsky.social
This week in #Tidytuesday I'm looking at top 5 ranked Pokemons in attack, defense and HP across generations. Defense and HP are way more stable than attack!

It was fun thinking how to work on the generations to make this plot.

code: github.com/mic-wypych/t...

#rstats #dataviz #ggplot2
A plot showing top 5 ranked pokemons in attack, defense and HP across generations. Each generation represents all generations up to this one (so e.g. generation 3 is generations 1, 2 and 3) which allows me to show which pokemons topped which in these 3 attributes. Defense and HP are much more stable than attack. In Attack generations form 1 to 4 introduced a new strongest Pokémon. In Defense since generation 2 Schuckle is strongest and since generation 3 regirock was 2nd and Steelix 3rd. In HP Since generation 2 Blissey is on top and Chansey is 2nd. Top HP Pokémons are dominated by normal type, top defense are more varied: bug, rock, steel and water. Top attack Pokémons are dragon, ground, normal and rock
April 3, 2025 at 9:30 AM
mwypych.bsky.social
It works. And I think sandwich supports clustered SEs at least for clm. It does not work for multinom though.
library(ordinal) library(marginaleffects) library(modelsummary) library(sandwich) library(nnet) library(MASS) wine <- ordinal::wine fit <- clm(rating ~ temp * contact, data = wine) #this works avg_slopes(fit, vcov = ~ judge, type = 'linear.predictor') #and this works modelsummary(fit, vcov = ~ judge) #and this works vcovCL(fit, culster = ~ judge) #trying multinom birthwt <- MASS::birthwt bwt.mu <- multinom(low ~ age + lwt + smoke, birthwt)
March 25, 2025 at 11:40 AM
mwypych.bsky.social
Also the code below actually worked for me so maybe at least clm will work?
library(ordinal) library(marginaleffects) wine <- ordinal::wine fit <- clm(rating ~ temp * contact, data = wine) avg_comparisons(fit, vcov = ~ judge)
March 25, 2025 at 11:03 AM
mwypych.bsky.social
Simple #tidytuesday week 5: it's donuts! 🍩
Donut chart of Simpsons: who spoke the most words.

#Dataviz in #rstats

code: github.com/mic-wypych/t...
A donut chart showing all members of Simpsons family and what percentage of all words spoken are by them. Homer: 42.29% Marge: 19.35% Bart: 17.49% Lisa: 17.1% Grampa: 3.57% Mona: .14% Virgil: .08% Maggie: 0% Marg
February 5, 2025 at 9:39 AM
mwypych.bsky.social
Today's accidents in #rstats #ggplot2

#dataviz
a gradient interval plot with coord_polar(theta="y") and a palette from purples through blue to teal a gradient interval plot with coord_polar() and a palette from purples through blue to teal
January 21, 2025 at 11:09 AM
mwypych.bsky.social
Really liked the week 46 #TidyTuesday data so I made another one, this time focusing on regime changes on various continents in each decade.

I was initally really excited about this but ultimately I feel it just shows how much missing data there is, especially in Africa

#rstats #Dataviz
A sankey chart showing changes in regimes types (communist, democracy, monarchy, interim and other/no data) on various continents/regions (Asia, Africa, Oceania, Americas, Europe) in each decade from 1950 to 2020. Data for all continents apart from Europe and Oceania has a lot of missing data.
December 12, 2024 at 1:21 PM
mwypych.bsky.social
Catching up on previous #TidyTuesday week 45
I looked at suffrage vs having women as a president in countries with presidential political system.

I think it's actually the first time I used {geomtextpath}!

#rstats #ggplot2

code: github.com/mic-wypych/t...
The plot shows number of countries having suffrage or not and number of countries having a women president or not in a given year from 1950 to 2020
November 12, 2024 at 12:08 PM