ParcMe

Free Parking
Team Name: 
NVM

Project Description

Parking is something that has infuriated us all at one time or another. There just never seem to be any free spots when you want them. But what if you knew the likelihood of finding a parking spot anywhere in the CBD for the exact time you wanted to park? What if you were given turn-by-turn directions for a route that maximised your chances of finding a park? We think there'd be less traffic, less frustration, and the world would be a better place.

ParcMe (Parking Availability in Real-time using Cox Modeling Engine... or maybe we'll rename it ParkMyRide) uses state of the art analytics and predictive modeling to tell you when and where parking spots will become available in your city. To do this, we use a combination of real-time data from parking sensors and a predictive model for telling us when car parks are likely to become free based on things like the time of day, the time each car has been in its spot, and the time each car is legally allowed to park in that spot.

With this real-time data and model, we can tell users the likelihood of finding a park in a particular place at a particular time, and even plot them a route that maximises their chances of finding a park within a particular area. Local governments can also use the data and model to forecast demand and plan infrastructure.

Prize Qualification

We think this idea has great potential as a startup. It's a hard problem, it's useful, and all of the hard stuff can be rolled into a dead simple interface that tells users where to go to get a park, or gives them simple metrics they can use to plan their travel.

Not only can we build a killer app and business, but we can also help relieve congestion and road rage. If people are informed about their chances of parking, many will decide to use alternative transport methods rather than take a high risk car journey. Furthermore, by giving people routing information that optimizes their chances of finding a park, our app can reduce traffic in already congested areas while people drive in circles looking for a park.

Data

Our model and real-time simulation uses Melbourne City's parking data for 2014. This is a very comprehensive data set, including every parking event occuring during the year, with information on arrival time, duration, departure time and location data. In all there are around 16 million records. Since the data set is so large, we've loaded it into a database and transformed it for easier processing. This allowed us to run queries, create graphs and most importantly, feed it into our duration model.

Project image fromĀ StockMonkeys.com

Datasets Used: 
Melbourne City parking event data for 2014: https://data.melbourne.vic.gov.au/Transport-Movement/Parking-Events-2014/mq3i-cbxd BCC parking meter data (cached version): http://webcache.googleusercontent.com/search?q=cache:MupwnnYvqokJ:data.brisbane.qld.gov.au/index.php/dataset/parking-meter-areas/+&cd=1&hl=en&ct=clnk&gl=au

Local Event Location: