Problem Statement

The real estate market poses challenges in accurate property valuation due to its dynamic nature and numerous influencing factors. Traditional methods, often manual appraisals, lack precision, scalability, and may result in inconsistent valuations. With the increasing demand for housing and limited land availability, there is a pressing need for efficient and data-driven approaches. This project addresses the problem of inaccurate and inconsistent property valuations by leveraging advanced machine learning techniques. The challenge lies in developing a model that not only incorporates a comprehensive set of features but also translates the complex relationships within real estate data into an accessible and user-friendly application for both professionals and consumers.

Objectives

The objective of this project is to develop a user-friendly mobile application integrating a pre-trained machine learning model using real-world housing price data. The app aims to provide accurate property valuation estimates, benefiting both real estate professionals and consumers. Through advanced machine learning algorithms, the project seeks to enhance prediction accuracy, offering valuable insights and facilitating more informed decision-making in the dynamic real estate market.

Research

The images below show how the research and exploration has been done in steps starting from the general user flow that can be done within the app. Once that has been done, we can go through the dataset to extract valuable information that the user has control of to be put as the input. Annotations play a huge role to ensure that all the information is displayed all the times. After we have carefully explored the data, we can start the designing process from the wireframe. Developing a wireframe is a crucial step since it is in low-fidelity, we can iterate multiple times to find the best and efficient solution.

Solution

The application facilitates user interaction by providing inputs correlated with key property attributes influencing house prices. Users can seamlessly input their preferences into the machine learning model, streamlining the valuation process. The interface is intentionally designed for simplicity, ensuring users can easily specify their housing preferences, contributing to a user-friendly and intuitive experience.

A tab of Buy and Sell has been placed in order to differentiate the type of inputs since in Buy, users do not need to fill in the exact value meanwhile in Sell, users will need to fill in the exact value to feed into the machine learning model. Thus, a spinner/dropdown menu is chosen to give user ease of use to fill in. Chips are also used to display all the facilities and nearby facilities that the user might want to have in their house. The use of emoji in the chips are also carefully chosen in order to give a more visual experience.

Once everything has been filled in the proceed button will activate to indicate the user that they can proceed to the price prediction process. The app will retrieve the value and feed it into the machine learning model as inputs to generate the output which is the house predicted price.

A loading page has been placed in order to give visualization to user that the model is currently running and in process. The illustration will be a GIF to give more interactivity to user and a more pleasant user experience in using the app.

In the result page, it distinctly displays the price prediction as well as the options that the user have selected from previously. This offers a conclusion on what the user wants in their house, and also so that they would not need to go back to see their previous choices. This removes an extra step and thus ease the navigation in using the app. Visual Hierarchy plays a huge role in this design in order to categorize things that is related to each other. This ensures the user of the app will have an easy time to navigate and utilize the app.

Remarks

This project is still in progress and hope to be done by this year. Stay tune for updates on this app :)