Inspiration

Imagine you are a student at Princeton University and your girlfriend lives you New York. You visit her frequently, but you are frustrated that the public transportation takes over around 2 hours while a car ride is only an hour. Every time you visit her, you schedule your leaving time to be the time that your friend Jim drives to New York and you carpool with him, or you ask around your friend group to see who's leaving for New York.

We want to simplify this process of asking around by centralizing all these information. To make it even simpler, our service makes it possible to request and post about rides without even going online.

What it does

TextRide is a service that intelligently recognize all your driving information through sending a SMS message. We will recognize the departing location and destination, your desired leaving time, and number of people for your ride through NLP and store it in our database. The service is only for university students for safety purposes, and it matches you up with other student drivers in the database leaving around the same time. If no direct match can be found, our system will suggest to you the closest destination, and propose a public transport option for the rider.

How we built it

We took in text messages and performed NLP, extracted key information for each posted/requested ride. We deployed on Google Cloud platform with Flask Framework. We stored all the info in PostgresSQL database. We used Standard Library and Twilio for sending SMS messages. We created web hooks with GitHub that performs CI/CD deployment.

Challenges we ran into

challenge #1: Current entity extraction libraries does not work specifically for extracting starting and ending locations, time to leave, and number of spots available automatically challenge #2: Using the google-maps-api to get geocode location and get best route challenge #3: parsing information from SMS messages to database and returning

Accomplishments that we're proud of

We have an app that is working and responding to people's messages. This service makes it very efficient to match riders with drivers, saving students tons of time and money. And it all works without requiring wifi or using data.

What we learned

How to use all the above mentioned technologies, especially interacting with SMS messages.

What's next for TextRide

Give riders screenshots of best google maps routes.

Share this project:

Updates