23.12.2022

Finding the best home at the right price with Data Science and AI

Bryan Bramaskara, Dr. Adrienne Heinrich, Gabriel Isaac Ramolete, and Dustin ReyesAI and Innovation Center of Excellence, Aboitiz Data Innovation

Property appraisal and value estimation are prone to human errors and bias due to price subjectivity and the lack of knowledge on the impact of surrounding amenities to the property’s value. Trustworthy sources on property information are limited and there’s usually a lack of access and transparency on the value of a property. These scenarios are observed in many countries including the Philippines.

Property value prediction is a modern approach to estimating property value. This method combines various data sources and advanced techniques to provide a more comprehensive and accurate prediction of a property’s value. By leveraging advanced technologies to combine and analyze data from different sources, such as property records, market trends, and local socio-economic indicators, it is possible to get a more complete picture of the factors that affect a property’s value. This can help to identify potential strengths and weaknesses of a property and provide a more accurate and fairer estimate of its value. Data science and Artificial Intelligence (DSAI) can play a crucial role in this process.

Leveraging Data Science and AI in property valuation

By applying DSAI to property value prediction, it is possible to analyze vast amounts of data, identify trends, and make more accurate predictions of a property’s value. There is clearly an opportunity for a fair property valuation service utilizing DSAI like machine learning with the aim of minimizing biases and providing reliable predictions.

In recent years, several use cases that employ machine learning in providing fair and unbiased property valuations were explored. However, these studies and services were mostly located in developed and highly established countries.

With this, our team at Aboitiz Data Innovation AI and Innovation Center of Excellence aim to create “Fair Homes,” an integrated service for finding the best home for customers’ individual needs at the right price. With the power of DSAI, customers can get an objective valuation of a property in the Philippines and countries with similar landscape/scenario.

Fair Homes: Integrated Property Valuation Service

The Fair Homes project aims to provide an accurate and reliable property valuation tool for customers, real estate agents, and appraisers. It targets to provide a decision support tool that can visualize the factors that would complement the valuation tool, such as hazard risks, amenities within the area, road patterns, and similar price trends, among others. Lastly, it intends to remove or minimize appraiser bias regarding property valuation.

Predictive models for property valuation routinely involve conventional features of the house/area, such as the number of bedrooms and bathrooms, floor and land area, and market prices of nearby properties.

Fair Homes targets to provide a data-driven, reliable, and fair property valuation prediction based on spatial property characteristics and socio-economic indicators. The integrated property valuation service suggests that alternative, non-traditional data should be incorporated to account for deviations in the true market value of a property and improve property value predictions in the Philippines and other countries with similar challenges while utilizing advanced analytics to leverage and visualize traditional sources of data. Government data, collected at the local government unit level, can provide quantifiable ways of ranking local government units which can depict the quality of living in the area. Amenities and buildings such as hospitals and schools will be included in mapping initiatives. Geohazards and accessibility factors will be considered as well. The project will also take advantage of satellite imaging geographic location, and street view images to extract visual features of the surrounding area.

Additionally, Fair Homes can offer customer services such as an interactive interface that can help potential buyers assess the livability of the location and property financing options matched with customer information or demographics. The service can also have a community platform for users where businesses, contractors, and home improvers are rated and referred, and where people can share home improvement recommendations with each other.

Considerations

One potential challenge of integrated property value prediction is that it can be time-consuming and resource-intensive. Gathering and analyzing data from multiple sources can be a complex and labor-intensive process. Additionally, the accuracy of the prediction can depend on the quality and reliability of the data used, so it is important to ensure that the data is up-to-date and relevant.

Another consideration is transparency and reliability. Responsible AI is of utmost importance for this service as a huge number of diverse stakeholders are intended to use this service. Having such a system that complies to the principles of responsible AI shall serve in providing fairness and explainability in a sense that decisions have basis.

To overcome these challenges, it is important to have access to high-quality data and the right tools to analyze it. This is where DSAI can play a crucial role, by providing the necessary tools and methods to extract insights and predictions from large and complex datasets. With the right data, tools, and expertise, it is possible to make more accurate and reliable predictions of a property’s value, and to provide valuable insights to property owners, lenders, and investors.

Impacts and opportunities

Overall, the integrated property value prediction offers a more comprehensive, accurate, and trustworthy way to estimate the value of a property. By combining traditional and alternative data sources and leveraging the latest advances in DSAI, it is possible to provide a more detailed, reliable, and fairer prediction of a property’s value, which can be useful for a wide range of purposes—whether you are a property owner looking to sell or refinance, or a lender or investor looking to assess the risk and potential return of a property. With our initial development with 20.41% Mean Average Percentage Error (MAPE) rate wherein the lower the better, we were able to compete with Hong Kong with 32-54% and other developing countries like Kuala Lumpur with up to 20.9% MAPE rates based on the publications that we have explored thus far.

This can also be among the considerations of the government for urban planning and smart city initiatives. Knowing the nearby amenities of the properties shall help property planners, buyers, and urban planners assess the quality of life within that area. The use of alternative data sources such as Cities and Municipalities Competitive Index in the estimation of property value can further motivate the government in investing in the competitiveness of their local government units based on certain facets such as economic dynamics, infrastructure, and resiliency. Apart from these, real estate planners can know in advance the potential value of a property knowing the current state of the area and its potential for growth through proxies such as point-of-interests, population growth, among others.

In summary, with more accurate, accessible, and transparent property valuations:

  • Buyers can reduce the risk of overpaying for a property;
  • Homeowners can sell their homes at the right price without misinformation;
  • Investors can determine the risk of purchasing property;
  • Potential buyers who are not naturally in the real estate network can get access to properties at almost the same time as members of those in the real estate circle; and
  • Appraisers and evaluators can build repeatable, accurate, and streamlined valuation methods to help inform their valuation practice.

Editors’ note: A paper on the “Utilization of Government-Based and Non-Conventional Indicators for Property Value Prediction in the Philippines” was published and presented at the 15th National Convention on Statistics organized by the Philippine Statistics Authority (PSA).

Related Papers