DataStorm's Webinar 2

Un outil pour anticiper les dérives du Machine Learning

R Package and app R Shiny used during DataStorm's webinar about MLOps.

Install and Run

In RStudio

Step 1 : Download data

The dataset used by the package can be downloaded at :

• https://www.kaggle.com/carrie1/ecommerce-data/home
• https://archive.ics.uci.edu/ml/datasets/Online+Retail

wget https://archive.ics.uci.edu/ml/machine-learning-databases/00502/online_retail_II.xlsx
mv online_retail_II.xlsx MLOpsMonitoring/inst/data/uk-retailer-ii.xlsx

Step 2 : Install R dependencies

options(repos = structure(c(CRAN = 'https://cloud.r-project.org')))
install.packages("remotes")
install.packages(c("DT", "MLmetrics", "caret", "changepoint", "data.table",
  "lubridate", "openxlsx", "rAmCharts", "shiny", "shinycssloaders",     
  "shinydashboard", "shinyjs", "stringr", "xgboost", 'randomForest', 'mccr',
  'Metrics'))

Step 3 : Install package

Step 4: Run app

library(MLOpsMonitoring)
MLOpsMonitoring::run_app()

With Docker

# Step 1 : Download package
git clone https://github.com/datastorm-open/webinar_mlops.git

# Step 2 : Build package with build.sh
bash build.sh

# Step 3 : Create Docker instance
docker build . -t datastorm_wbnr_mlops

# Step 4 : Run Docker instance
docker run -d -p 3838:3838 datastorm_wbnr_mlops