20.1 early beginnings
In 2016, I published my Master thesis work on high intensity rainfall depending on temperature,
see rainfall use case.
I also started a PhD in flood research and was part of a taskforce to analyze a flashflood in Braunsbach (southwestern Germany), see
nearbyStations use case.
For both these projects, I developed some R code to download and read data for given weather stations. In June 2016, these scripts got included in my misc package berryFunctions, see the commit history there.
The DWD had just started their CDC FTP server with a serious amount of data the year before, so this was an exciting time to work on it.
rdwd started life as a standalone package in October 2016.
Back then, it only had dataDWD and readDWD. That quickly expanded, see the
early rdwd commit history.
In January 2017, it was first published on CRAN.
A bit of development was done in early 2017, then I quit my PhD (apparently, software dev is not valued much by grant-acquiring research institutions) and spent 2018 mostly offline building a house.
20.2 a giant leap…
Early 2019 marked a turning point: I went to France for three months and worked a lot on
rdwd, introducing support for reading asc and binary raster files.
That was outsourced to
dwdradar over summer to keep the
rdwd development version easy to install (i.e. without FORTRAN code).
Systematic testing and raster plotting/projection methods got implemented and documentation took a leap forward.
Over 2019, an increasing number of feature requests spurred further development, including creating this homepage (now I hardly ever need to answer “yes, it’s already implemented, look here” anymore ^^).
A lot of automation went into updating the indexes of available files and runnning all local tests (code and index). Nowadays, I just need to trigger them by calling the two related functions.
2020 (especially the summer) then saw a lot of refinement in details, slowly morphing
rdwd into a mature and stable package.
More detailed (but still aggregated) changes can nicely be seen at https://github.com/brry/rdwd/releases
20.3 the future
I plan to continue maintaining the package, even though its capabilities have long exceeded my personal needs.
Coding simply brings joy - and of course it’s also very satisfactory to see my work actually used in many contexts.
One big thing I have planned is an interactive app to visualize the weather of a given time period, compared to the measurements of the same period in other years.
A few (rather minor) issues are also still open and I expect that state of things to continue for a long time :).
Lastly, I hope to find some help in understanding the structure of gridded data to improve that part of the package.