Testing epishiny

[adamkucharski]

Key links

Link to GitHub repository

Dependencies

# install.packages("remotes")
remotes::install_github("epicentre-msf/epishiny")

Need to install pkgload (not declared correctly as a dep)

Initial problems identified

Package documentation not fully loaded (e.g. via roxygen2) because can’t find key functions after installing from GitHub.

Potential reason: don’t have @export tag in all functions. Maybe need devtools check?

Work around

1. Clone repository
2. Source files from set up

library(epishiny)

# example package data
data("df_ll")
data("sf_yem")

# etc...

3. Load functions locally

file_path <- "~/Documents/GitHub/epishiny/R/"
source(paste0(file_path,"01_time.R"))
source(paste0(file_path,"02_place.R"))
source(paste0(file_path,"03_person.R"))
source(paste0(file_path,"04_filter.R"))
source(paste0(file_path,"05_launch.R"))
source(paste0(file_path,"utils.R"))
source(paste0(file_path,"utils.echart.R"))

4. Run test version

launch_demo_dashboard()

Date filters work. Group filters seem to have longer to load.

Potential solutions for day 2:

1. Load locally and focus on epidemiological features. Some potential options:

• additional modules (R estimation, cfr package etc.)
• alternative line list data (e.g. from cfr case studies
• feedback on visualisation options and other feedback.
• explore transmission chain module to examine use of multidimensional data (e.g. via existing highcharter functionality).

2. Update code so correctly loads as package. But perhaps better for Paul at Epicentre to have first pass at this, unless quick and obvious fix.

[henryls1]

Potential solutions could include landscaping of future module developments

[henryls1]

###Other data sources can include:

1. global.health EVD Uganda (2022)
2. global.health MVD Equatiorial Guinea (2023)

[Mougk]

Can also add mpox line list

[adamkucharski]

We have summarised activites in the ‘for review’ issues here: https://github.com/epiverse-trace/epishiny/issues (we were working off fork for ease of sharing, but will move key issues/PRs to main repo once decided).

Good ones to start with to get idea of app structure would be:

• Document structure of app
• Test app on other datasets
• Document new modules

Then 'Add tables' and 'include options' summarise specific functionality that is going to be submitted as draft PR

[JamesABaker]

Thanks! Looks like there has been a lot of work going on! 🐝

I've added some comments to the issues you suggested, but this is obviously just a cursory glance at what looks like an incredibly complex problem! Please feel free to question and challenge what I have said 😄

[adamkucharski]

Thanks for the feedback. Summary of output PRs here:
epicentre-msf/epishiny#9 (comment)

[ntncmch]

For those interested in being involved in the development of epishiny, please add a 👍 to this post so I can add you to the meeting I'll organise with Paul Campbell (@PaulC91), our R Shiny dev, in the coming week or two.