Back to Projects

Team Name:

Square


Team Members:


Evidence of Work

Coral - protect the lives

Project Info

Team Name


Square


Team Members


yuyang wang , Chongzheng Zhao , Ayaz , Harshal Shah and 1 other member with an unpublished profile.

Project Description


Team Member -
Ayaz Aziz Mujawar
Harshal Harish Shah
Chongzheng Zhao
Yuyang Wang

Awards Categories Entered
Proactively reducing rubbish and pollution in our waterways
Track and trace: help end plastic waste

Problem
- Proactively reducing the rubbish environment.
- Reducing the effect on aquatic life.
- Targeting Hotspot areas(suburbs)based on the collection of waste.
- management of garbage so water health can be improved.

Solution
- Create an application named "Coral" that helps to detect waste in real-time using Convolution Neural Network.

  • Also It helps to understand which suburbs are more prone to wastes based on different parameters like events(such as Footy games, Big Bash cricket matches, etc), climate, and location.

  • Forecasting the waste collection for the next 5 years(until 2025) for City of Melbourne.

Technical Details -

Back-end API development
Using Flask as a framework as an API interface to the coordinate data displayed on the map.
Use a lightweight development base on the back end and add features as needed
The back-end API is connected to a Mysql database and has a certain error tolerance.
The back-end developed a cache mechanism, because the query coordinates of the data requires algorithmic computing time, so the cache can greatly reduce the length of time the user waits for data display
Coordinate processing back-end development
The camera locations are not around the river, so to simulate a polluted river and make the project more meaningful, all camera locations are automatically assigned by the algorithm to a specific river location.
The calculation is done using a matching search algorithm based on the morhattan distance.
The use of cache mechanism greatly reduces user waiting time
Front-end Dashboard Development
The front-end theme framework using Wordpress for secondary development
Use responsive pages on the front end and integrate modules such as data display, map display, video display, project monitoring, etc.

Front-end Maps Development
We use the Google Maps platform, which uses real-time data and real-time imagery to create an immersive location experience that allows users to make more appropriate business decisions. The main advantage of Google Maps is that it allows users to embed "maps" into a website without having to build their own map server. By bringing in the Victoria River Map dataset into the map, users can clearly see the distribution of rivers in Victoria. In addition, data from individual monitors is presented as markers on the map, while the data of monitors are matched to the rivers nearest to them based on an algorithm, allowing a clear view of the waste data collected by the nearby monitors integrated on each river. The garbage data is also color-coded depending on the amount. In addition, we also used Echarts, which is an open source visualization library implemented in JavaScript. It relies on the vector graphics library ZRender, which is both intuitive and interactive, as well as highly customizable data visualization charts. After retrieving the data from the server, the data is further formatted and processed.

Front-end Dashboard Development
- The front-end theme framework using Wordpress for secondary development
- Use responsive pages on the front end and integrate modules such as data display, map display, video display, project monitoring, etc.

Convolution Neural Network
- Used PyTorch for CNN and pretrained the data using ResNet architecture.

Forecasting
- Used the ARIMA Time Series model for Forecasting.

Tableau - Tableau Dashboard is used for creating the correlation visualization between the City of Melbourne suburbs and events conducted in the suburbs.

Correlation
https://public.tableau.com/profile/ayaz.mujawar#!/vizhome/Dashboard1_15975371903980/Dashboard1

Forecasting
https://public.tableau.com/profile/ayaz.mujawar#!/vizhome/Dashboard_15975369465540/Dashboard2

Application Dashboard
Please navigate to http://govhack.eventplus.cc/

[NOTE: IF MAP DOESN'T WORK IN NORMAL MODE PLEASE OPEN THE WEBSITE IN INCOGNITO MODE]


Data Story


Our team has used data from different open source websites like Melbourne water's Hotspot data and geojson data. We have used data from "litter stoppers email data", "City of Melbourne open data", "Bureau of Statistics" data. Along with that we have taken waste image dataset from one of the github site.


Evidence of Work

Video

Homepage

Team DataSets

Melbourne Water Dataset

Description of Use This dataset we used to calculate the latitude and longitude of the rivers in the City of Melbourne area.

Data Set

Melbourne Water Dataset

Description of Use Used for detecting the waste volumes in the hotspot areas

Data Set

Waste Image Dataset

Description of Use Image datasets for Convolution Neural Network.

Data Set

Environment Waste collected

Description of Use Dataset about environment waste collected

Data Set

Beaureu of Statistics

Description of Use Data about seasons like rainy, windy, etc

Data Set

Dataset about events

Description of Use Dataset about events

Data Set

Litter Stop Data

Description of Use For accessing the Latitude and Longitude of the litter regions in Melbourne and also to analyze the waste images from dataset.

Data Set

Melbourne Water Dataset

Description of Use For detecting and comparing the values of target Hotspots.

Data Set

Challenge Entries

Proactively reducing rubbish and pollution in our waterways

When rubbish and pollutants enter our waterways, it harms our aquatic life and spoils the amenity of our community spaces. How can we use open data to proactively reduce rubbish and common pollutants from entering our waterways so that we protect our natural environment and keep our rivers and creeks clean today, so that we can enjoy them tomorrow?

Eligibility: Must use at least one Melbourne Water dataset.

Go to Challenge | 6 teams have entered this challenge.