Team Name:
NextGen
Team Members:
NextGen
Problem Statement:
Victoria continues to be the fastest-growing state in Australia, with our population projected to hit 10.3 million by 2051. The state's economy grew by 9.1% over the past two years, surpassing growth rates in NSW, Queensland, Western Australia, and Tasmania. In terms of business investment, Victoria has seen an 11.3% increase as of December 2023, significantly higher than the national average of 8.2% (Invest Australia, 2024).
.jpeg)
As we face high growth pressures from factors like population change, natural increase, net overseas migration, and interstate migration, the demand for housing, roads, and infrastructure is escalating. To address these needs effectively, we need to adopt more predictive approaches for urban planning. Our challenge is to forecast how community dynamics—such as housing demand, population trends, and traffic patterns—will evolve, enabling planners to make proactive and informed decisions for sustainable development.
.png)
Project Description:
Our solution is a web-based application designed to assist urban planners, policymakers, and community stakeholders in forecasting the future dynamics of communities across Victoria. Powered by AI (Artificial intelligence) and ML (machine learning), the app leverages historical data to predict trends in population growth, housing demand, vehicle registration, building permits, and public service needs. By using Victorian government datasets, our solution aligns with real-world planning goals, enabling users to explore various scenarios and evaluate the long-term impacts of urban planning decisions. In future, this can be combined with live databases to ensure real-time tracking of urban development plans in the state of Victoria.
By providing data-driven insights, we ensure that planners can balance population growth, creating livable, inclusive communities for future generations.
Goals:
1. Observed Trends Data on past planning permits will give insight into the historical trends of distinctive residential and commercial growth and transformation, land-use changes and infrastructural developments.
2. Empower Planners by forecasting Future Changes: Train AI models to predict future shifts in population density, demand for housing, infrastructure needs, and demand for public services.
3. Present Predicted Urban Futures: Develop a predictive visualization tool that lets stakeholders experience various urban futures with different variables.
Mission:
We want to help Victoria build liveable, sustainable and equitable communities by creating the best possible predictive model. The goal is to learn from past trends to develop more sustainable urban areas with predictive AI-powered web application.
Data Story:
Our project focuses on predicting community evolution within Victoria by using AI and historical planning datasets. The analysis utilized various datasets, such as traffic count locations, population density, vehicle registrations, housing development, and Infrastrcture premits. The goal is to forecast changes in urban dynamics, traffic, and population density in Victoria, enabling better urban planning and resource allocation for sustainable development.
Data Collection:
Data Transformation:
Data transformation: Cleaning and normalizing raw data from various sources is a critical step in creating informative data. This involves cleaning, normalizing and finding connections across categories between various datasets. For example, traffic data (Traffic flow, Geo coordinates, Average annual daily traffic, Intersection description) is merged with population density data and historical urbanization trends to make a combined dataset.
Forecasting:
Utilizing machine-learning modelling, our project will generate analyses of predictions for how the communities in Victoria will likely develop to meet the increasing demands of future generations. The project will use predictive machine-learning models to anticipate historic changes in population growth, in vehicle registrations, traffic patterns, and housing development. Such predictive models generate detailed maps showing ‘hotspots’ of traffic, as well as anticipating where housing communities might expand next. The anticipated insights will enable planners and policy-makers to use machine-learning models to direct resources and responsibilities to build sustainable urban development, maintain transportation systems and better manage public services, to meet increasing demands of future generations. We have utilized two forecasting models among various other methods as mentioned below.
ARIMA Model (AutoRegressive Integrated Moving Average): ARIMA is a popular statistical model used for time series forecasting. It combines three components: autoregression (AR), differencing (I) to make the data stationary, and a moving average (MA). It works well for short-term forecasting and when the data shows strong autocorrelations. ARIMA models require the data to be stationary, meaning the mean and variance should not change over time.
Equation:
ARIMA(p,d,q) = AR(p) + I(d) + MA(q)
Prophet Model: Prophet is a forecasting model developed by Facebook, designed to handle seasonality, trends, and holidays with flexibility. It is particularly effective for time series that contain strong seasonal effects and are influenced by special events (e.g., holidays). Prophet is user-friendly and works well with missing data or outliers. It also doesn’t require the data to be stationary, making it easier to use compared to ARIMA.
Equation:
y(t) = trend(t) + seasonality(t) + holidays(t) + ε(t)
Technical Implementation
Our app is built using the following data science libraries and tools:
With datasets like traffic patterns, Leaflet can plot real-time data about where traffic is increasing or where congestion is expected. This is vital for forecasting future traffic patterns and identifying high-traffic zones.
The Challenge and Goals:
The goal is to develop a model that can predict and visualize the evolution of community behaviors to support urban planning on key issues such as housing demand, traffic problems, administrative boundaries and service provision decisions. The primary challenge we face is identifying the most accurate and comprehensive dataset to create an effective predictive model. While historical planning permit data is valuable, critical real-time data, such as housing market trends and traffic patterns, is often unavailable in time series format.
The lack of this data restricts our ability to capture dynamic changes that directly impact community development, such as fluctuating housing prices, shifts in traffic flow at different times of the day or year, and immediate demands for public services.
To mitigate this challenge, we focused on what we have and how that data and informaton can be used to create something poweful and efficient enough to drive meaningful results. For Example, area of Gelong is used in housing predictions because of uavalibity of other datsets, which provided reference to how more models can be created for future, making it as work in progress project.
Strategic Deployment:
Our model will be user-friendly, and able to plug into planning tools. This will allow city planners to simulate cause-and-effect relationships across the system, and generate actionable insight for more foresighted urban planning.
Conclusion:
By predicting coming community dynamics, the project will enable Victoria to plan its urban future with greater foresight, sustainability and equitability. This is accomplished by using past data to take into consideration present and future planning decisions, as well as to provide citizens with knowledge on the long-term outcomes of urban decisions.
With the massive changes in residential and commercial growth of Victoria, knowing the time frame from which historical data can be extracted will be a useful tool in forecasting.
https://www.abs.gov.au/statistics/people/population/regional-population/2022-23#data-downloads
https://www.abs.gov.au/census/find-census-data/historical
https://www.localgovernment.vic.gov.au/elections-2024
https://discover.data.vic.gov.au/dataset/housing-development-data-2013-2019
https://www.vic.gov.au/department-of-transport-and-planning-future-directions
https://discover.data.vic.gov.au/dataset/traffic-count-locations
Newly registered vehcles in Victoris time series dataset from 1900
https://discover.data.vic.gov.au/dataset/monthly-new-vehicle-transfers
Housing Development Data 2013 - 2019 - Victorian Government Data Directory
knowledge-for-sustainable-australian-cities-and-communities.pdf (futureearth.org.au)
https://www.mdpi.com/2413-8851/5/3/68
https://www.planning.vic.gov.au/guides-and-resources/Data-spatial-and-insights/discover-and-access-planning-open-data/urban-development-program/victoria-in-futureType the / key to add custom blocks
Description of Use We used this dataset for forecasting by applying time series analysis to identify population growth trends across different regions. Predictive models can help estimate future population changes at various geographical levels, aiding policy planning and resource allocation.
Description of Use This dataset is used to forecast how many new vehicles will be registered in Victoria based upon the fuel type.
Description of Use We are leveraging this dataset to model Victoria's population growth through demographic components, enabling government authorities to implement data-driven urban planning strategies and resource management for future development.
Description of Use We are utilizing this dataset to forecast Victoria's population using time series analysis, enabling government authorities to make data-driven decisions for urban planning and resource allocation in the near future.
Description of Use Defined and accurate spatial boundaries are essential to useful spatial analysis. They deliver regionally focused delivery of population changes, housing need assessments and public service demand management. At the same time, they provide a basis for geographically definitive insights and predictions.
Description of Use This dataset is used to forecast future houses development in the Gelong area of Melbourne, act as a basis for other predictions.
Description of Use Analyzing building permits is important for understanding the pace and direction of urban growth as it provide information about the housing market’s trajectory, indicating potential supply and demand imbalances. . It also helps in forecasting the need for utilities and public services like roads, schools, and waste management systems.
Description of Use The datasets are used to predict high-traffic spots in Victoria by analyzing the AADT values with X and Y corrdinates. With this data, city planners and traffic management teams can identify congested intersections and midblock roads, helping in decision-making for road improvements, traffic signal placements, and reducing congestion. The year column help in assessing how recently the data was collected, ensuring up-to-date traffic predictions.
Go to Challenge | 13 teams have entered this challenge.
Go to Challenge | 26 teams have entered this challenge.
Go to Challenge | 16 teams have entered this challenge.
Go to Challenge | 35 teams have entered this challenge.
