Home

OnePark Documentation

Problem Statement

With Singapore moving towards a Smart Nation, our carparks are become more outdated by the day. There exist information gaps where residents know about the state of the carparks but not authorities, and authorities know all the available parking lots but not the residents.

Value Proposition

OnePark serves as a platform for information to be exchanged between residents and authorities (HDB) in a fun and rewarding manner.

Features

To entice drivers to use OnePark, we have added in additional features to make it an all-in-one carpark app. It shows the current available parking lots, predict the number of parking lots available within the next 24 hours, and even a search option to find the next nearest carpark with parking lots.

OnePark also includes a voice assistant, so that road users can receive assistance while driving without the need to look at their phone screens, thereby ensuring that road safety is not compromised.

OnePark offers an open platform for fault reporting. All faults reported will be reflected in a simple user interface for all users to see.

OnePark draws on the success of the National Steps Challenge, where we would incentivise carpark users to take on an active role in the upkeep of HDB carparks. This comes in the form of a goodwill point system, of which the points can be used to redeem for “Scratch&Win” cards, petrol vouchers and more. To earn points, carpark users will have to show public spiritedness in the form of proper and accurate fault reporting, maintaining good user rating through good parking practices, etc. Users’ rating will be affected in cases of ill-usage of the carparks.

Tech Stack used

React (Webpage) Express (Webpage) SQL (Database) Python (Machine Learning)

Unique selling point – Lot Predictive Model

OnePark uses Machine Learning technology to provide an accurate 24-hour estimate of carpark lot availability. Our model is trained on a dataset retrieved from Carpark-Availability-Data API from https://data.gov.sg/dataset/carpark-availability. We decided to use the XGBoost library for the XGBRegressor, training it on features such as “Area Code”, “Total Lots”, “Carpark Lots”, “Hour of Day”, etc to predict the carpark’s lot availability at a given time.

Inspiration

Presently, carpark users do not adopt good parking practices because there is little incentive to do so. Hence, we have decided to come up with OnePark to encourage all users to take on an active role in the upkeep of our carparks. We were primarily inspired from the success of the National Steps Challenge to help Singaporeans break away from their sedentary lifestyle.

What it does

OnePark provides all-in-one services: from open fault-reporting to lot availability and maintenance predictions using advanced Machine Learning, to help residents take responsibility for their estate.

How we built it

Users access our application from their mobile phones at our frontend server running ReactJS. Before our frontend server displays the map, current incident reports as well as user rewards available, it communicates with the backend which processes these queries with the MySQL server and returns the data required.

The backend is also responsible for machine learning and modelling by running the algorithm used for carpark lot prediction when a user makes a request on the application.

Challenges we ran into

As this was our first time building a web programme, it took a lot of time to coordinate all our efforts together but we are thankful that we succeeded in the end.

Accomplishments that we're proud of

We managed to create a working web programme with most of its features built in within the 48 hours.

What we learned

1) Putting an interactive google map onto the webpage.

2) Using Machine Learning to predict Time Series Data

What's next for OnePark

The Machine Learning model can predict the estimated lot availability of the carpark at relatively high accuracy. Following this success, we would be able to predict estimated time taken for a fault to arise, to notify relevant authorities to pre-emptively send the necessary manpower to check for any faults before any occurs.

As the take-up rate increases, OnePark can be a one-stop portal for all carpark users, for payment of parking fees, checking of parking lots, fault reporting and as a liaison between the relevant agencies and the general carpark users.