Inspiration

Admit it! Finding Parking is hard. Specially, in Old and Downtown Louisville, or any major cities in the world. According to the data collected from 2014, Louisville streets have been operating at 120 percent to 130 percent of capacity. According to dailycamera.com, The Louisville Downtown Street Faire, on nine Friday nights in the summer, now regularly draws several thousand people to the city looking for room to park. Although there are apps that tell you how much parking spot is available they are not real time, and by the time you get to a parking spot it might have already been taken.

What it does

SmartPark will obtain real-time parking lot data using Google maps API to find the parking lots in the areas and use users location at repeated intervals to calculate the number of vehicles that have occupied a parking spot. Not only do we calculate the empty spaces available but also provide the user with a reasonable estimation of the parking spots being available. The approximation only gets better with more user using our app since we will be able to improve the parameters for our model.

How we built it

We built it using Python, HTML, JavaScript, API, Google Maps API We then use Poisson Distribution to approximate the probability that the empty spaces are going to be occupied in a time interval The parameters for our model (rate of cars looking for a parking space) using a Mixture of Gaussians model.

Challenges I ran into

Biggest challenge was to properly estimate probability distribution.

Accomplishments that I'm proud of

Create a probabilistic estimation model from scratch

What I learned

Big Data Analysis, Statistics, Web Development

What's next for SmartPark

Right now we are simulating the data for our model parameter, but we would like to obtain real training data to approximate our parameter.

Share this project:

Updates