Skip to content

rNLKJA/Australia-Social-Media-Analytics-on-the-Cloud

Repository files navigation

2023 S1 Cluster and Cloud Computing
Australia Social Media Analytics on the Cloud
Social Sence Dashboard

The project focuses on an export of Twitter data from the Australian Data Observatory ADO , data to be harvested by students from the Mastodon APIs and data from the Spatial Urban Data Observatory SUDO. The focus of this assignment is to use a large Twitter corpus (that will be provided) to tell interesting stories of life in Australian and importantly how social media data can be used alongside/compared with/augment the official data available within the SUDO platform to improve our knowledge of life in Australia.

Team 57

Name Student ID Email
Zongchao Xie 1174047 [email protected]
Xuan Wang 1329456 [email protected]
Runqiu Fei 1166093 [email protected]
Wei Zhao 1118649 [email protected]
Sunchuangyu Huang 1118472 [email protected]

Directories

From listing folder, there are five parts of our project. Therefore:

  1. For Backend Application, access 1_Flask_Backend
  2. For Frontend Application, access 2_ReactJS_frontend
  3. For CouchDB deployment, access 3_CouchDB_database
  4. For Data Processing and Data Scraping, access 4_Python_data_processing
  5. For Ansible IT Automation, access 5_Ansible_IT_Automation

See README in each folder for detailed implementations.

Dependencies

  • Backend: Python3>=3.9, Flask
  • Frontend: Javascript, Node.js>=12, React.js
  • Database: CouchDB v3.2.1
  • Data Scraping: jupyterlab>=3.0.0, Python, MPI
  • IT Automation: Ansible, YAML

Instance Access

Documentation

Data Sources

Scenarios

  1. Examine the distribution of tweets discussing income inequality, financial struggles, or job satisfaction across different income brackets. Identify geolocations with higher median incomes and compare the sentiment of tweets in these areas to those with lower median incomes.
  2. Analyze the relationship between criminal incidents by principal offence and the frequency of tweets discussing crime or safety in specific geolocations. Investigate how the sentiment of these tweets varies across areas with different crime rates.

Special Thanks

License

The code will be public after 27th May 2023. For @copyright information please refer to MIT License.

Team 57