Smart Park Analyst

Inspiration

Seeing students and staff have issues parking on campus, especially during busy times for commuters arriving on campus. We also have friends who use disability parking and are constantly having issues getting the parking spots they need to be able to get around on campus. Lastly we also know people who try to find when charging locations on campus are available.

What it does

The app allows you to view live counts for total parking, disabled parking, and charging parking spots available at any given time. It can click parking lots on the map and view the live stats. It allows the ability to input a building or parking lot and it will predict what lot you should park at and when to ensure you get the desired parking spot. It has a community feature allowing people to mark that they have parked or left the parking lots which is especially useful for the special parking spots.

How we built it

We are using a Django backend with an SQLite database that manages data and returns data to the frontend. We created some example data for trends and usage of parking on campus for 12 weeks of "parking usage" and trained models using sklearn to have the functionality to predict the availability of total spots, disabled parking spots, and charging spots for any parking lot for any given time. The frontend is built with react bootstrap with the openstreet map API with the react leaflet library to create maps on the frontend.

Challenges we ran into

We initially tried to use MongoDB for the backend but the package Djongo which was meant to translate the SQL generated by Django ORM had many shortfalls so we pivoted back to SQLite, in production we would likely use Postgres or MySQL. Issues with the Map API and showing the data that we wanted for the application.

Accomplishments that we're proud of

Proud that we came up with a good idea. Proud that we got the map API working. Proud that the API models are working with 97% for regular parking spaces, and 84% for disabled and electric parking spots. Proud to have learned Django for the backend.

What we learned

The Django framework, The OpenMaps API.

What's next for Smart Park Analyst

We would want to integrate the live data from the scanners of the parking lots. We would want to integrate data from the ChargePoint API to know which spots are open or will be open soon. We would want to capture as much data as we can to enhance our prediction models making the product more effective with more time.