Overview of Open Source Work

This is a guide to my github repo. I have a lot of projects and it could be difficult to get an overview. Although most of my professional work is with python, most of my github contributions are in R.

• Showcase projects

• Packages

• Datasets

• [Scraping and machine learning](#Scraping and machine learning)

• [Shiny apps](#Shiny apps)

• outside contributions

Showcase

Running R scripts on a schedule (invertedushape bot)

All my blogposts about scheduling can be found under 'scheduling'. Overview page that gives advice on what to think about when you want to run automated R scripts. And a blogpost that describes the options can be found here. To demonstrate this I created a bot that generates a random plot and tweets this to twitter. The actual bot is stupid, but the main point is that this is an example of scheduling an R script, passing secrets, and making sure the package versions are pinned.

Actual repository with code and explanation for heroku, github Actions and Gitlab Runners and a version packaged into a docker container.

tools: R, heroku, cron, azure, github, gitlab, docker APIs: twitter

Using dbt with postgres on github

I've used the standard dbt example of jaffle shop and changed some details so the results are displayed in github pages. code repository

tools: dbt, python, postgresql, docker, github

Quering wikidata and tweeting the result

This is a small demonstration based on the basic wikiData query example where we retrieve the people born on this date but using the date of death. Packaged into a docker cotnainer. tweetwikideath tools: R, heroku, cron, docker APIs: twitter, wikidata

code examples

Setting up a poor man's CI/CD for shiny apps from github Actions

blogpost, actual code

tools: R, github, docker APIs: shinyapps.io

How to use TidyModels on UbiOps

This sets up a preprocess-training pipeline with a trained model as result. and a a preprocess-predcit pipeline to predict new results. blogpost repository with code.

tools: R APIs: UbiOps

Code to create XKCD like plot of cities

https://github.com/RMHogervorst/xkcd_weather_cities_nl Inspired by XKCD. We can determine our ideal city by looking at wintertemperature and humidex (combination of humidity and summer heat). This is a very silly thing to do for the Netherlands, because the country is so small and flat that the weather is approximately the same everywhere.

tools: R

Code to add dancing bananas to any picture

http://rmhogervorst.nl/cleancode/blog/2017/11/28/content/post/2017-11-28-building-the-oomsifier/ https://github.com/RMHogervorst/bananafy The code uses the magic package through the magick wrapper in R. It is a simple script that turns

tools: R, magick

Code to float a small piece of country away

https://github.com/RMHogervorst/floating_friesland

tools: R

Packages

PinboardR

](https://www.repostatus.org/#active)

https://github.com/RMHogervorst/pinboardr https://rmhogervorst.nl/pinboardr/ A package to talk to the pinboard API. You can use this package to add new bookmarks, delete them, extract all bookmarks, or to add, modify or remove tags. Basically everything you can do through the website, but from an R session. This is a full implementation of all the endpoints in pinboard.in/api.

Gephi

](https://www.repostatus.org/#active)

https://github.com/RMHogervorst/gephi https://rmhogervorst.nl/gephi This is a simple package to export files into a csv format that gephi can understand. This package does not interface with the open source network vizualisation software gephi but it writes and reads in the same csv format as gephi.

I’ve found the need to convert tidygraph/igraph objects into a node and edge csv, to visualize in gephi quite often. This should be trivial, but gephi is a bit particular and wants specific column names.

Forensicdatatoolkit

](https://www.repostatus.org/#active)

https://github.com/RMHogervorst/forensicdatatoolkit

These tools merely tell you if the underlying data is probable or not. It can never tell us why the statistics are wrong. Someone could lie or someone could mistype a number. Only open data can tell us which it is.

banana

](https://www.repostatus.org/#active)

https://github.com/RMHogervorst/banana An exploration of what can be done with the {vctrs} package. This is an implementation of an S3 vector which displays sizes in relative sizes (bananas) but keeps the precise value in cm underneath. This allows you to calculate with the real values but keep the display in relative values.

IMDB

https://github.com/RMHogervorst/imdb https://rmhogervorst.nl/imdb/articles/walktrough.html The package imdb helps you in downloading series and movie information from imdb (it uses the omdbi api). It has three functions one for basic information about series and a second one that also downloads synopsis, actors etc. A third function downloads information about movies.

gpodder

https://github.com/RMHogervorst/gpodder

musicgraph

connect to musicgraph api. another api https://github.com/RMHogervorst/musicgraph

A package for house style in coffee colors

https://github.com/RMHogervorst/coffeegeeks This package contains basic needs for every rstatscoffeegeek.

• an rmarkdowndocument template with coffee colors and image
• a ggplot2 coffee theme
• basic help for when your coffee doesn't taste great
• coffee related emojis
• coffee names

Datasets

Werewolf dataset

fictional data set of three classes https://github.com/RMHogervorst/werewolf I needed a dataset with many interesting features, but mostly where I can probabilistically assign people to three categories.

The people in this dataset can be divided into three categories. They are either normal, werewolf, or wererabbit. All 'normal' measures such as haircolor, BMI, eye color, etnicity, BFI values, blood types, favorite food and allergies match the global values of the USA, Age is a random birthday sample. The names come from the babynames package and are selected from the years between 1960 and 1993. So the general values will be equal to the US population. However I only adjusted the hair and eye values for race, other things are not conditional.

Unicorns on unicycles dataset

A practice dataset for data munging practice. https://github.com/RMHogervorst/unicorns_on_unicycles

Coffee drinking moments

one of my first packages. https://github.com/RMHogervorst/coffeedata

Star trek

I scraped moviescripts from star trek the next generation and turn that into a dataset. I'm not even that much of a fan but the data was available.

TNG

TNG dataset, contains every line and every description of all the episodes of TNG https://github.com/RTrek/TNG

Speak random sentance from character

https://github.com/RTrek/startrekpackage This package gives you the option to talk like the main characters of star trek The next generation. Call the

Scraping and machine learning

Code to scrape the gdpr fines website

https://github.com/RMHogervorst/scrape_gdpr_fines I extracted this dataset for tidytuesday.

Code to extract and parse notes from kindle

Kindle highlights are saved into a text file on the device. If you export that you can pull it apart into a useful dataformat: repository and explanation .

Code to download and extract transcripts of a certain podcast.

Security now podcast. https://github.com/RMHogervorst/NLP_SN

• downloading the podcast transcripts
• extraction of relevant information from the transcripts

Shiny apps

Karaoke song selector

https://github.com/Raoke/powrballad https://rmhogervorst.shinyapps.io/powrballad/

Contributions

I made a small contribution to apache airflow. apache/airflow#8240