Language: ENG 🇬🇧 | 中文 🇨🇳

China Unicorn Index

#R #Shiny #Neo4j #Docker

Background

An application built on R + Shiny which integrates with Neo4j graph database and shipped in Docker. Data is extracted from Hurun Report (English | Chinese), which published an index of China unicorns (company with more than 10 billions in valuation).

The purpose of this exercise is to provide a demonstration of connecting R to graph database, and packaging exploratory tool into a viable data product.

Setup

Build Docker Image

git clone https://github.com/bj-sodas/china-unicorn-index.git
cd china-unicorn-index
## Note: Replace CRAN mirror url (https://cran.r-project.org/mirrors.html) with the one that is closest to you. ---
docker build -t china-unicorn-index .

Or pull image from Github package registry

docker pull docker.pkg.github.com/bj-sodas/china-unicorn-index/china-unicorn-index:latest

Configuration

We are required to create a config file named config.yml in your directory. It is used for Shiny app to locate Neo4j database and authentication information. A sample is provided in this repository.

default:
    title:      # Something that displays on application title bar
    url:        # Neo4j database and port e.g http://172.30.10.9:7474
    username:   # Neo4j username
    password:   # Neo4j password
    inode:      # Initial node to display

Running Application

The following provides 2 methods to run the application. Recommended way of doing it is through docker-compose (method I) as services are defined for you. Otherwise, you can choose to run the application separately (method II).

(I) Run App ( with Docker Compose )

We will define and run 2 services, one for our Neo4j database (named neo4j-db) and one for our Shiny app (named webapp).

First we extract the database snapshot to Graph folder so that data can be pre-populated.

tar -xvzf graph.data.tar.gz -C Graph

Then we specify the configuration as the following,

default:
    title: China Unicorn Index
    url: http://neo4j-db:7474 # docker will resolve address of neo4j
    username: neo4j
    password: somepassword
    inode: Ant Financial

On command line run

# make sure you are in 'china-unicorn-index' base directory
docker-compose up -d

Open your web browser and locate http://localhost:33838/china-unicorn-index.

(II) Run App ( from local runApp )

Start a Neo4j Docker (named neo4j-db) container with the following options,

• ports 7474 (HTTP) and 7687 (Bolt) exposed;
• binds the import directory (so that we can import data through csv files);
• create username (neo4j) and password (somepassword) for authentication;

# make sure you are in 'china-unicorn-index' base directory
docker run \
    --name neo4j-db \
    -p7474:7474 -p7687:7687 \
    -d \
    -v $PWD/import:/var/lib/neo4j/import \
    --env NEO4J_AUTH=neo4j/somepassword \
    neo4j:3.4.0

With Docker up and running, we import our data by passing queries through bin/cypher-shell.

cat Graph/setup.cql |  docker exec --interactive neo4j-db bin/cypher-shell -u neo4j -p somepassword

Database is ready. Here we specify the configuration as the following,

default:
    title: China Unicorn Index
    url: http://127.0.0.1:7474 # neo4j is exposed to localhost
    username: neo4j
    password: somepassword
    inode: Ant Financial

Run the application in RStudio, or on command line run

R -e 'shiny::runApp(host="127.0.0.1", port=33838)'

Note: You must have required packages installed in your local station.

Open your web browser and locate http://localhost:33838.

FAQ

My application cannot start. It says An error has occurred. Check your logs or contact the app author for clarification.

• Check your config file. Have you declared your url correctly?
• Give neo4j some time to start up, and refresh your browser.