Inspiration

Experiencing first hand having trouble finding a parking spot on busy days and knowing others have too.

What it does

Lets the app user know which parking spots are available for the selected parking lot.

How we built it

A mix of HTML / PHP (Web Server), Dart (Flutter, used for making the mobile app), and Python (Deep Learning for training the model).

Challenges we ran into

Not enough data sets, not enough time to process the given data sets, the model not generalizing very well due to small data sets, problems with apache webserver restrictions (prevented the PHP from writing to a text file, fixed later), time zones (not everyone was on at the same time)

Accomplishments that we're proud of

Being able to view the model even though it's not perfect (86% accuracy), dabbling in new programming languages and learning the basics, being able to combine programming languages into one result.

What we learned

Where to modify apache write permissions, how to send HTTP Post requests from Dart and Python, how to manage data sets, training a CNN model to be able to recognize images, how to navigate through RowdyHacks.

What's next for Park-Or-Not

Increasing the amount of data sets for better accuracy, adding more parking lots to analyze, and adding design to the app interface. Improving the over all accuracy and efficiency of the model.

Built With

Share this project:

Updates