In HF-Research/HTData: Provides data for HjerteTal

HjerteTal

HjerteTal (HT) consists of four main components:

1. DST project 706322

2. github repository HF-Research/HTData

3. github repository HF-Research/ht2

4. hjertetal.dk

Here we describe the workflow from the raw register data to the public shiny app which needs to be run at every update.

1. DST project 706322: At the moment it is places in the folder V:\Data\Workdata\706322\MariannaMeaidi\HJERTETAL_V2.\ For now we have only updated the numbers for cardiovascular disease, and not congenital heart defects (the next step in the future). Both projects are constructed using the {targets} package. (see https://books.ropensci.org/targets/{.uri}). The definitions of all the outcomes (diseases, operations and medications) are kept in CSV files in the ‘input/’ directory, called ‘diag_ht2_table_ny, ‘behandling_table’, and ‘med_table’ respectively. These tables contain the ICD/SKS/ATC definition, along with the patient type (ambulatory/admitted) and diagnosis type. Additionally, all outcomes are assigned a “HjerteTal code” or identifier that I use to refer to them in the R code (to avoid writing out the full name every time).

The export files that we use in the Shiny app are kept in TXT files in the ‘output/’ directory. One file for each disease-group, one file for the operations and one file for the medications. There is one file for the output combos aswell.

When your programming is done and validated, you need to export the export files out of DST and into GitHub.\

The final product - The Shiny App - is maintained on GitHub, which can be found here: https://github.com/HF-Research/ht2. (Clone it to your local machine to get started) . In the HTData folder you will find the export data that will be used in the Shiny App. If you are unfamiliar with Shiny, please start with this: https://mastering-shiny.org/. There is lots and LOTS of tutorials online if you get stuck. This app is developed and tested on a local machine before it is deployed to shinyapps.io where it is hosted to be viable to the public. Both the server code and the ui code are broken into separate files, one for each “page” of the app (technically, one for each ‘NavbarPage’). There are 2 projects that are needed to run the Shiny app:

• GitHub repository HF-Research/HTData: The R package that holds all HjeteTal’s data exported from DST.

  • Dump the .txt files exported from DST into the ‘inst/extdata’ directory

  • Run the R scripts found in ‘data-raw’ directory (Note, these need to be modified/added to if there are changes to the format in which data from DST is exported)

  • Re-build the package and push the changes (R-studio -> Build -> more -> build source package). Remember to change the version number in the file DESCRIPTION.

  • Install the package –> Install –> Install Package

• Github repository HF-Research/ht2: The project that contains all the Shiny code. To update data in HjeteTal (for example, when a new year of data become available), the data is first put into a separate package called HTData. The process for updating the data for this package is:

• In the ‘ht2’ projekct, re-run the data pre-processing.R script. Check if the newest version of HTData is loaded: packageVersion('HTData')

• Check that everything works by running locally by running global.R and run runApp()

• Check that everything works on the server by deploying to ‘ht-dev’ (see the ‘deploy.R’ script)

• Deploy to the production (see the same ‘deploy.R’ script)

PROJECT FOLDER STRUCTURE IN HT2:

Below is a description of each folder in the shiny project

• www/ contains various icon files and JS files for doing things like: checking if the user is using Internet Explorer, Google analytics code – note this is associated with the hf_forsking Google account

• ui/ contains the shiny code for the UI. Each navBar page has its own script, and language specific objects are kept in their own language ui script.

• server/ contains the code for the server logic. Like the UI code, each navBar page has its own script.

• rsconnect/ This folder is automatically generate and updated when deploying – do not manually edit

• R/ contains all the functions that are used in both the server and ui scripts

• pre-processing/ contains the data-pre-processing script. All the other scripts are not used

• language/ contains data files made during the pre-processing step. These are made to reduce the time on start-up of the app for the end-user.

• input_ui_text/ contains the CSV files that control the text show to the user in both English and Danish. Not all the columns in each file are used, some were added, but never ended up in the final version

o variable_ui_chd: contains the text for the description of the events for congenital heart diseases (prevalence, incidence, etc)

o variable_ui: contain contains the text for the description of the events for acquired heart diseases (prevalence, incidence, etc)

o update_descriptions_chd: describes any updated we make to the CHD database. This is important as sometimes new data, or new methods mean that the numbers might change, even for year back. This can cause worry among users sometimes, so it is nice to explain why things change

o update_descriptions: Same as above, but for acquired CVDs

o ui_about_text_chd: contains the text for the ‘About’ or ‘Vejledning’ section o ui_about_text: same as above but for acquired CVDS o outcome_descriptions_chd: contain the name, and description of the outcomes (ie the diseases, procedures, or medications).

o outcome_descriptions: same as above but for acquired CVDs

o faq_chd: contains text for the FAQ table for CHDs

o faq: same as above but for acquired CVDs o ethnicity: contains description for how ethnicity is defined in HjerteTal

o edu_description: contains description for how education groups are defined in HjerteTal • inti/ contains the code for deploying the app to shinyapps.io

• documentation/ I believe nothing in here is used anywhere

• data/ contains various data file that are use by the data-preprocessing step (i.e. geographic data for the maps, valid_output_combos (see 1.1.2 for a description of that file). Files that are .rda or .fst file are datafiles created during the data-preprocessing step. This folder is a bit of a mess, and there are probably item that could be deleted

THINGS TO NOTE When you deploy to shinyapps.io, the shiny app will be running on a Unix machine. Because this is different than the Windows machine you will presumably be developing the app on, this can sometimes cause bugs. One example, the way windows encodes non-latin characters is different than Unix (I spent A LOT of time on this!), so when using Danish-specific letters, it can be tricky. That’s why you often see ‘enc2utf8()’ surrounding Danish text. I do not entirely understand any of this...

HF-Research/HTData documentation built on April 3, 2025, 2:18 p.m.