SEAMLESS DASHBOARDING WITH SHINY + FLEXDASHBOARD + HEROKU
What is LAG-RUG-DASH
LAG-RUG-DASH uses Flexdashboard and Shiny, which allow for rapid web
application development as dashboard by capitalising on the usage of
R-markdowns. This helps to attain the final stage of the Data Science
workflow or pipeline which is communication of analysed data.
<hr>
</hr>
Dependecies
- rmarkdown
- httr
- scales
- flexdashboard
- shiny
- DT
- plotly
- naniar
- devtools
Getting Started
To install LAG-RUG-DASH from source you need Rstudio and R lang 3.4.4>
in order to get started. Run the following in command prompt
git clone https://github.com/tosi-n/LAG-RUG-DASH.git
cd LAG-RUG-DASH
then open the project folder LAG-RUG-DASH from Rstudio
File Paths
Alternatively
To install LAG-RUG-DASH from google drive link, after download Rstudio
and R lang 3.4.4>. Download all from link
below
https://drive.google.com/drive/folders/1jOiaKovigCXNB16HI5v9SJQIkwBm-bR4?usp=sharing
then open the project folder LAG-RUG-DASH from Rstudio
Install Dependencies
In Rstudio, install package dependencies using the init.r file. Open
init.r and click run to install deependencies on Rstudio.
Design and Layout
Layout operates in 4 TIERS based on client or user requirements
- Sidebar layout (Input {.sidebar}) SB
- Tab layout T(Tab-name)(======) T
- Rows and Columns (Row{}) R
- Widgets / Components
C
Edit code in LAG_RUG_DASH_v0.2.Rmd to fit your requirement. <br></br>
Note: sidebar is used for creating dashboard input data or feature to be
implemented on the entire dashboard.
Hosting on Heroku
In order to have our LAG-RUG-DASH Livee on heroku, follow the below
listed requirements and process.
init.R
Procfile
Aptfile
run.R
heroku-buildpack-r
As indicated earlier, the init.R file contains packages to be installed
on heroku server. Heroku dynos uses a contanerized feature to host
application. For heroku dynos to know the kind of application we are
install we need to set Procfile as web feature so that heroku know our
application is a web application. Procfile also carries information
about the file root directory and the console type to
use.
web: env fakechroot fakeroot chroot /app/.root /bin/sh -c 'cd /app && /usr/bin/R -f /app/run.R --gui-none --no-save'
As seen above the Procfile requires a set of Linux console environment
in order for it to operate. Using the Aptfile, we tell heroku to install
the below listed Linux console in order for heroku to run R and
dashboard packages in the server.
libpq-dev
libxml2-dev
gsl-bin
libgsl0-dev
With the above done, R shiny is the runtime set for the dashboard to
operate on so therefore, a run.R file is needed for the dashboard
rendering. The run.R file is used to:
- Set port for heroku to listen to
- Set the default working directory
- Run the dashboard on the server heroku creates
<!-- end list -->
port <- Sys.getenv('PORT')
setwd('/app')
rmarkdown::run(
'LAG_RUG_DASH_v0.2.Rmd'
, shiny_args = list(
host = '0.0.0.0',
port = as.numeric(port)
)
)
git push Heroku Deploy
Create heroku account, then go to heroku dashboard and create
application by clicking on New. Set application name and set server
location as US or Europe then create.
After creating a heroku account and application, go to Rstudio terminal
run the following commands:
heroku login
git init
git add .
git commit -m "commit message"
heroku git:remote -a app-name
heroku buildpacks:add http://github.com/virtualstaticvoid/heroku-buildpack-r.git#heroku-16
heroku stack:set heroku-16
git push heroku master
Check deployment success using command in Rstudio terminal:
heroku logs --tail
Update Deployed Dashboard
For constant update to dashboard on a live heroku server, go to Rstudio
terminal run the following commands:
heroku login
git add .
git commit -m "commit message"
heroku git:remote -a app-name
git push heroku master
