Project Description
A collection depot has million of trash and some of them already had an object detection technology to filtering but they haven't used that data to make an application for the community yet.
The more precisely people trash in, the less effort for delivering and collection depots also.
Imagine that we have thousands of waste which are considered bottles such as juice
bottle, water bottle, and sports water bottle. A question arises is whether all bottles can be recycled, while the information
about recycling on trash bin is so limited. Technologies we have used are 3D scanning and AR. Firstly, Environment Protection Authorities can scan numbers
of waste and upload to server then the mobile app downloads that data. While a user is confused about their trash, they open
the camera in the mobile app to scan that object. The result will appear as a sticker on that object to indicate its type and the correct bin to put in.
We make a mobile application that not only educates people living in Australia about waste filtering but help them with detecting trash types to throw in the correct bin also.
Moreover, we encourage everyone to keep their city and neighbourhood clean by recording their correct trashing. At the end of the day, they will get some small rewards for good conduct.
Data Story
We use the data of landfill waste, green waste and recycling waste in the city of Ballarat, we combined these 3 data sets into 1 table and sum up daily data into monthly data. We made descriptive analysis to show the amount of waste in the 3 categories in the past 3 years, then we use ARIMA model to make prediction of the amount of waste will be generated in the coming years. We found the recycling waste in the future will be unstable like it is now, we conclude that one possible reason is people are not using the recycling bins correctly. Thus we came up with the idea of making an mobile APP to help people to know what kind of waste should go to recycling bin. In the APP, it also reward people who collect more recycling waste, we believe this APP will help to solve the problem of the recycling waste not being put in the right spot.
1.Clean Data
select c.date,a.greenwaste,b.landfillwaste,c.recyclewaste from
(select date,count(vehicle) greencount, sum(weight) greenwaste from [kerbside-green-waste] group by date) a right outer join
(select date,count(vehicle) landfillcount, sum(weight) landfillwaste from [kerbside-landfill] group by date)b on a.date=b.date right outer join
(select date,count(vehicle) recyclecount, sum(weight) recycle_waste from [kerbside-recycling] group by date)c on a.date=c.date WHERE C.DATE BETWEEN '2016-07-03' AND '2019-06-30' ORDER BY 1;
This table combines the original 3 datasets
Ballarat Kerbside Recycling waste Deliveries
Ballarat Kerbside Green waste Deliveries
Ballarat Kerbside Landfill Deliveries
We combined 3 tables into 1 table and sum up daily data into monthly data.
2.Data Statistical description
This shows the average monthly waste in 3 categories, min, max and std deviation.
Line charts give a brief idea about the amount of waste (unit: ton) generated in the recent 3 years.
3.The monthly deviation plot with Recycling Waste
4.Apply Seasonal ARIMA model to forcast the data
5.Forecast of Recycling Waste in the city of Ballarat