dclaze / big-data-trend-visualizer

an implementation of using hadoop to process big data and visualize it in a web ui

big-data-trend-visualizer

an implementation of using hadoop to process big data and visualize it in a web ui. our first planned visualization is to use Nasdaq ITCH data transformed in VWAP. See 'Contribute to VWAP Visualization' below.

Demo: /web

Design

API

Deploy

Prior to running deploy steps, please complete 'Prerequisites'

1. cd api
2. export AWS_PROFILE=big-data-trends-visualizer
3. mvn clean install package -Drelease.version=prod && serverless deploy --stage=prod

Prerequisites

1. Install nvm, see https://github.com/nvm-sh/nvm
2. Install node, run nvm use v12
3. Install serverless, run npm install -g serverless
4. Install serverless plugins npm install -g serverless-plugin-scripts
5. Install Maven, see http://maven.apache.org/guides/getting-started/maven-in-five-minutes.html
6. Install AWS CLI, see https://docs.aws.amazon.com/cli/latest/userguide/install-cliv2.html

eg. on Mac I ran brew install awscli

7. Run aws configure, and enter your AWS credentials. The aws_access_key_id and aws_secret_access_key are secret, reach out to get access.

Edit ~/.aws/credentials and ensure the the profile is named big-data-trends-visualizer

 [big-data-trends-visualizer]
 aws_access_key_id = ********************
 aws_secret_access_key = ****************************************

Web

The web front is composed of 2 views web/home/index.html and web/visualizer/index.html

Run Locally

1. cd web
2. npm install
3. npm start
4. Open URL http://127.0.0.1:1643/home/ in web browser

Adding dependencies

1. cd web
2. npm install
3. npx bower install <package-name>, see https://bower.io/ for more
4. IMPORTANT: We are using github pages to host the webpage, please check in all bower_components files

VWAP Visualization

Our planned prototype will target a VWAP Visualization. The idea is to create a visualization for the Nasdaq ITCH data and generate a Volume Weighted Average Price visualization.

1. Add front end visualization to web/visualizer/index.html

Use D3, see https://d3js.org/, or other library

2. Split NASDAQ data

Determine the schema of the NASDAQ dataset, see S3 bucket, see https://console.aws.amazon.com/s3/buckets/nasdaq-itch/?region=us-east-1

3. Write Hadoop Map Reduce Algorithm

Transform from ITCH to VWAP on a cluster of Hadoop machines

4. Connect the input and output of Hadoop result to the API

Integrate our on-deman API with the Hadoop cluster. This requires us to deliver the input and output to the Hadoop cluster, store it somewhere to be retrieved, and wire it up to an API that returns the result to the customer

Hadoop

To see the Hadoop VWAP implementation, see /hadoop. This code is responsible for converting from ITCH-5.0 to a data trend for a given stock symbol. To see more details on how we convert from Nasdaq ITCH see data/nasdaq/README.md.

The main hadoop application can be found at hadoop/src/main/java/dk/njit/cs643/itch50/vwap/VWAP.java

Resources

• Google Drive Assets