Real Estate Appraisal Software
Value Estimate Factors
- COUNTRY UAE
- MAIN TECHNOLOGY Python
- OTHER TECHNOLOGIES Laravel, React.js
Investing in real estate is a tough nut to crack, especially in highly competitive markets, like Dubai, for example. It is challenging to understand which investment can bring expected returns and when. Not to mention, comparing different offers can be a frustrating task.
Vetted Property is an online platform that addresses exactly these challenges. It is an interactive tool that empowers buyers and sellers during the crucial decision making process related to investing in real estate. It is a data-driven, customizable, user-friendly, and AI-based tool that is all about leveraging the Automated Valuation Model (AVM) to find investments with higher ROI.
How it works
Vetted Property is a web portal and a combination of different tools and technologies that support real estate investment processes. At the core of the solution sits a custom-built Automated Valuation Model that provides instant estimates of the market value for a selected property. The power of this AVM is that it not only does the general maths using Big Data and Artificial Intelligence but also calculates potential uncertainty levels related to a given investment and explains what can speed up a more detailed price estimation process.
Apart from the AVM, Vetted Property includes an online platform that supports investors in real estate with plenty of relevant and valuable investment metrics to consider. There is also a handy investment calculator that helps to dig deeper into the actual costs of each investment, shedding some light on important aspects such as cash flow or calculating expected ROI. All that is accessible in a convenient subscription-based model for individuals and organizations.
Design & UX
We were contacted by the client to improve their design and make it more user-friendly. We spent a week with the client, doing different UX workshops together, understanding their needs, and preparing several UX deliverables such as the product vision board, the user personas, the customer journey map, the impacting map, and the user story mapping.
In the meantime, the UI designer started preparing the app’s full sitemap, adding every step and page. The next step was to prepare a wireframe and select several benchmarks from other competitors. The final step was to design the high-fidelity prototype and a basic design system.
The final design is an elegant, uncluttered and professional layout, suggesting an innovative and tech industry, with rounded buttons and a fresh font style. We added different secondary colors based on the four main sections (investor, broker, developer, and lender), and the blue tone follows the client’s color. The single pages are methodically ordered with content, divided with accurate gaps and white spaces. The client preferred illustration graphics instead of photos, and we provided them with a new set of modern and 3d illustrations. The final result highlights the content for briefly explaining a new complex technology that can change the real estate industry in the following years.
The cooperation started with a Product Discovery Workshop, during which we prepared the product vision board, created personas, and mapped their user journey. We mapped user stories and impacted each scenario. Also, we focused on the design and functionality of the web portal, analyzed data and designed the AVM, and set up the infrastructure based on the cloud.
We used PHP with Laravel framework and React.JS to build the web portal. We used MySQL as a database and AWS as a cloud service provider. Pytorch and Python were used to build the AVM and implement neural networks, Shap was used to explain predictions. Also, we used Pandas for data wrangling, which means cleaning and unifying complex data sets in preparation for analysis.
Apart from that, we integrated the tool with payments provider Telr, integrated the Web Portal with AVM, and enhanced it with a Selenium-based automated data scraping solution that is crawling the web to find official sources on data related to analyzed properties.
More than ten people worked together for six months to deliver the platform and AVM. We had regular sprint reviews, plenty of spontaneous extra meetings, and everyday Slack conversations to ensure we stayed on course.