Back to Projects

Team Name:

AquaDash


Team Members:


Evidence of Work

AquaDash

Project Info

AquaDash thumbnail

Team Name


AquaDash


Team Members


Noah , Haard , Olga , Olaf Wrieden

Project Description


The issue of sustainable water use presents a growing challenge in Australia, driven by the increasing consumption, climate variability and low reuse. AquaDash is a water tank recommendation tool that quantifies potential water bill savings and provides personalised guidelines on what tank the user needs, based just on their address. AquaDash will empower the Sunshine Coast residents to drive more efficient, sustainable water usage for their households and for the region as a whole.

AquaDash uses Artificial Intelligence to estimate the user’s roof surface area from the Google Maps image and combine that with the water tank size guidelines to recommend a water tank size for the household. Using the rainfall data for their location in the Sunshine Coast region, the tool estimates how much water can be collected and saved by adopting the recommended water tank. By combining this information with the user household’s water consumption data from Unitywater, AquaDash quantifies and presents the potential savings the user can achieve in a personalised dashboard.

Tackling the problem statement

The project addresses multiple problem statement elements:

Weather and climate events and their impact on water use and infrastructure

  • The recent drought season is a stark reminder that the scarcity of water presents an excruciating challenge for many rural Australian households, even with today’s water infrastructure capabilities. AquaDash can equip households with actionable information that can help their household become more resilient to climate extremes and anomalies.

Water savings, analysis and reporting tools

  • AquaDash allows the user to view their water usage history and potential water bill savings based on the recommended tank size, rainfall data and UnityWater’s tier pricing information. If required, there is a future opportunity to present this weekly.
  • An interested customer simply supplies one input (an address) and AquaDash intelligently calculated the building’s roof surface area, the annual rainfall, and how much rainwater could be captured by the roof. These insights are then presented in an accessible manner via charts and text, informing the user at every step of the way. Including, how many times an average washing machine could be run from the rainwater that can be collected. The report culminates in a tank recommendation for the property, including a direct link to the council regulation.

Learnings from tracking and monitoring water consumption by geographic profile

  • By linking the Managed Object ID to the household, we are able to provide an accurate overview of historical water consumption for the household and forecast the comparative savings that would come from the water tank. There is a future opportunity to utilise the postcode and suburb information to segment the water usage insights by location.
  • The use of geographic data can also help drive valuable insights to the user, such as how their consumption compares to the suburb’s average.

Looking forward

Below is a summary of the ways in which AquaDash can be extended to provide an even better experience to the community:

  • Using the ManagedObjectid-linked geographic data to give the user insights about water consumption in their area and how their household compares. By going further and exploring the use of more fine-grained household address data, we could create a social experience where the users can see how much water they are using/saving compared to their neighbours/friends on the opt-in basis, given the appropriate data privacy considerations. This would gamify the experience of consuming water, with friends aiming to meet certain consumption reduction targets (seasonally adjusted). This could potentially also be further incentivised with water discounts to be used each quarter.

  • Expanding the savings information to include more household appliances that could be powered by the collected water. Allow users to register their devices (and associated energy ratings) to deduce:

    • More accurate estimations for how much the user could save by implementing a tank.
    • Track consumption changes after adding this device to their home (e.g. show the effect of different devices on consumption)

#water #sustainability #harvesting

Data Story


AquaDash utilises multiple datasets and Machine Learning techniques that can be divided into three key steps:

  • Roof surface area calculation and tank size estimation: AquaDash leverages the Google Maps API to decode addresses into latitude and longitude. From here, we once again use it to capture a satellite image for the property used to calculate the roof's surface area. Once the surface area is exacted, we colour it in blue and overlay it over the image. This is what becomes visible to the end user as the "calculated surface area".
  • Monthly rainfall estimation for a given locale: Using data from the Bureau of Meteorology (BOM), we were able to source a 36month average of rainfall across all stations in Australia. We divided this by three to get a stable estimate for annual rainfall. Our API then uses this data to locate the 3 closest stations to the address entered by the user, to triangulate an estimated value for annual rainfall in the area the house is in. We then augment this by the BOM monthly average rainfall data to deduce the estimated monthly rainfall in the region. Ideally, we would source this data at a monthly level to begin with, but the approximation was made due to reduced data access/ time constraints.
  • Monthly water usage for a given Managed Object ID: We save the UnityWater consumption data in a storage account, and then load it into an Azure Data Explorer cluster. We the Kusto Query Language (KQL) to process the daily Unity water datasets and extract, for a given managedObjectId, the monthly water consumption for that managedObjectId. We account for both digital and integrated water meters (summation over Pulse1 typeM values for digital and /10266/0 for integrated).

Evidence of Work

Video

Homepage

Project Image

Team DataSets

Google Maps API

Description of Use AquaDash leverages the Google Maps API to decode addresses into latitude and longitude. From here, we once again use it to capture a satellite image for the property used to calculate the roof's surface area. Once the surface area is extracted, we colour it in blue and overlay it over the image. This is what becomes visible to the end user as the "calculated surface area".

Data Set

Water efficiency

Description of Use We got the washing machine water consumption value (3-star rating) to use for savings recommendations

Data Set

Water Account Australia

Description of Use Used for general understanding of the context of the challenge and the problem of water sustainability in Australia, such as water consumption by state and cost for households vs industry.

Data Set

Unitywater - Digital Water Meter data

Description of Use The datasets were used to extract monthly water usage for a given Managed Object ID. We used Azure Data Explorer’s Kusto Query Language (KQL) to process the daily Unity water datasets and extract, for a given object ID, the monthly water consumption for both digital and integrated water meters (summation over Pulse1 typeM values for digital and /10266/0 for integrated).

Data Set

36 Month Rainfall Data per weather station - BOM

Description of Use We use this to get the rainfall average over the last 36months for the 3 closest stations to the address given. This helps us approximate what the rainfall for any given Australian house would be over that period. This is used downstream to get estimates for monthly rainfall, based on the rainfall trends published here: http://www.bom.gov.au/climate/averages/tables/cw_066062.shtml Assumptions made: - Australia monthly rainfall trend can be applied regionally. - Average rainfall for the last 36 months can be an accurate forecasting metric for future rainfall - The 3 closest stations can approximate any given addresses observed rainfall accurately

Data Set

Challenge Entries

Integrate Disparate Data Sources like a Palantir Engineer

When working with our customers worldwide, our most successful outcomes come from our ability to make sense out of high degrees of complexity. We think that leveraging our existing toolkits and performing data fusion in your projects will help you better solve the problems of GovHack 2022.

Eligibility: Fuse multiple sources of data into a single pane of glass (such as graphs or pivot tables); and/or leverage the Blueprint UI Toolkit in some or all of your designs and prototypes.

Go to Challenge | 9 teams have entered this challenge.

Moreton Bay greening as we grow

How might we harness the power of the everyday citizen to help protect our diverse flora and fauna as we grow our region; creating a diverse and flourishing planet for generations to come?

Eligibility: Open to everyone. Employees of Moreton Bay Regional Council with a direct working relationship with members of the local Moreton Bay GovHack node organising committee are ineligible to apply for this prize. If unsure, please feel welcome to check-in on Slack and discuss via the #hack-moreton-bay and #talk-mbrc channels upon commencement of the event.

Go to Challenge | 9 teams have entered this challenge.

Digitising your drinking water

Clean water is an essential service for our communities and has been identified as the ‘Blue Gold’ for the next generation. How might data be harnessed to influence people to reduce water consumption? How might monitoring and tracking water usage now help protect this valuable resource for future generations?

Eligibility: Open to everyone. Must use the Unitywater dataset. Must use at least one other dataset.

Go to Challenge | 11 teams have entered this challenge.

Reducing climate impact through sustainable energy behaviours

How might we reduce climate impact through changes in energy sources, production, distribution and consumption

Go to Challenge | 10 teams have entered this challenge.