README.md

In bbc/insights-latent-feature-explorer: insights-latent-feature-explorer

insights-latent-feature-explorer

Produces an app that explores the item and user matrices output by our recommendation model

Overview

• Project title: insights-latent-feature-explorer
• Inception Date: September 2019
• Main Stakeholder department(s): A list of interested departments (e.g. sports/childrens/etc)
• Main Stakeholder(s): A list of interested parties
• Author(s): Joshua Feldman

Details

How was the project initiated?

..Please complete..

What was the goal of the project

..Please complete..

Outputs

..Add any more info here if required..

See vignettes/

Usage

Setup

The R ecosystem is slightly different to that of Python and the approach to virtual environments is different. On 2019-08-02 an introduction to renv was created by the folks at RStudio. For the purposes of this template, it's recommended that we also use renv.

To learn more about how renv works see here.

To install the latest renv install the following in your system-level version of R.

Note: check that R will use the proxy if on the BBC network.

file.edit('~/.Renviron') # in Rstudio, or in any text editor of your choice

then add to the file and save / write out:

https_proxy=http://www-cache.reith.bbc.co.uk:80
http_proxy=http://www-cache.reith.bbc.co.uk:80

Now you should be able to do the following:

..If anything more than R package install required, use docker..

1. Install docker desktop - https://www.docker.com/products/docker-desktop
2. Run make docker - this will build and then run the container, will take a while the first time
3. Instructions to access via a terminal or web interface will be printed to the console Note: see thoughts / issues on the use of Docker to manage dependencies here. For more general information on R environments and Docker see here.

..Otherwise..

Option A: Using make (the easy way):

1. (Optional) initialise a renv for the project and activate:

make create_environment

1. To make sure your environment is stored run make package before checking in

2. There are other things you can easily run using make:

Run make help to remind yourself:

> make help:

clean               Delete all temp files (including environment) 
create_environment  Set up R interpreter environment 
data                Make Dataset 
docker              Run docker container, will build it first if required 
docker_stop         Stop the running container 
make shiny          Build and run shiny app 
package             Package dependencies to lockfile 
sync_data_from_s3   Download Data from S3 
sync_data_to_s3     Upload Data to S3 
test                Run tests inside renv 
update_environment  Sync Dependencies with Lockfile 

Tests

Run the unit tests using make test

Data

..If you are making use of s3 to back up data..

Data is stored here: s3://[OPTIONAL] your-bucket-for-syncing-data (do not include 's3://')

To get data: 1. Set up AWS cli on your machine, download the cli from AWS and then run something like the below on an EC2 box to generate some credentials to use: instanceid=$( \ curl -s 'http://169.254.169.254/latest/meta-data/iam/security-credentials/' ) curl "http://169.254.169.254/latest/meta-data/iam/security-credentials/${instanceid}" 2. Once access is granted you can run make sync_data_from_s3 to download data to use

Running

..Include instructions on how to run your program, or get up and running with your package.

Results

..Any more info on visualisations or where results might be saved. Under the package structure currently in use, we suggest using vignettes.

bbc/insights-latent-feature-explorer documentation built on Nov. 3, 2019, 2:08 p.m.