Mutad

The name Mutad is a reverse spelling of datum.

Overview

An implementation of a real-time data platform/search engine based on various technology stack. The implementation is rather for my own learning purpose.

<img src="./img/heatmap.gif?raw=true">

The above gif image demonstrates a real-time visualization of the tweet geoparsing of Mutad. According to my proceeding research, the amount of tweets with geospatial metadata is quite a few, say 10 per 10,000 tweets, from the Hosebird data feed. Thus I took another strategy to extract geospatial data from the tweet. Mutad implements real-time geoparsing based on CLAVIN. Because of my poor implementation, the accuracy is not that high yet.

<img src="./img/timeline.gif?raw=true">

The above gif image demonstrates a list timeline displaying the latest Tweets ordered from newest to oldest with a specified search query.

<img src="./img/trend.gif?raw=true">

The above gif image demonstrates a trend chart displaying the top most trended hashtags and tweet counts per a minute in the last hour.

Architecture

<img src="./img/mutad.jpeg?raw=true">

This diagram illustrates the overall architecture of Mutad. The technology stack is as follows:

• Apache Storm
• Apache Kafka
• CLAVIN
• My TRIE
• Elasticsearch
• Spring Boot
• React
• Next.js

Getting Started

Launch services by following the instructions below:

• Set up ELK Environment
• Set up Kafka Cluster
• Set up Storm Cluster
• Build Storm TRIE library
• Deploy Storm Collectors Topology
• Deploy Storm Processors Topology
• Start Spring Boot Elasticsearch Backend
• Start React User Interface

Future Work

The project is still WIP. See Issues.