The inspiration for this application definitely came from the movie "The Big Short" as well as some real life experience myself. I had some roommates in college last year whom had taken out student loans but weren't getting jobs to pay off those debts. Not many people are well versed in the language of loans but everyone understands money. We wanted to help students frame their decisions relative to the amount they will owe out of college. For example, most students who take, say a $100,000 dollar loan from a private lender and don't start paying it forward in college could end up with monthly payments of about $2,200 after they graduate. Overall, this comes out to about 1/3 of one's entire post graduation salary over the year. This is an incredibly amount, but by just paying some of the balance forward each month, you can help reduce these drastic payments and still avoid defaulting or becoming delinquent on your student loan. Student loans are also the first loans that people will be taking out in their life, most likely and if their credit scores are wrecked by becoming delinquent or defaulting, this could haunt them for the rest of their lives. will help you plan for yodebt.

What it does

The app aimed to integrate a seamless chatbot experience with the calculation of a student loan payment. Many people don't make an effort to get to know much about their loans because it's all numbers but we wanted to humanize the process a little bit more. The chatbot would take information such as where you go to school, your school's rank, the location, your major, and other metrics to help determine the margin interest rate. Combined with the average treasury rate in the US, and you have a base interest rate that resembles that which might be given by a private or government loan. Not many people know that their interest rates ARE indeed determined by those metrics as a bank needs to gauge how likely you are to pay back on the loan. It would then port those pieces of information, from Dialogflow into Google's firebase database. The Spring MVC we created would then request the information from the database and execute the mortgage calculations code, and be able to tell the students exactly what their monthly payments are going to be and how they can start paying it off now, and if they did start now, what would be the difference in their overall post-graduation payments.

How we built it

We first had to complete the mortgage calculator that Aman had created. From there, we started to train the Dialogflow chatbot. After that we started to look up how to port the outcomes from

Challenges we ran into

A big challenge we ran into was training Dialogflow to ask the right questions. Usually chatbots rely on the users asking questions and the bot answers them as best as it can. However, we wanted the chatbot to be prompting information from the user, and then storing the responses in a database. The database wasn't being formatted perfectly when we were trying to send the information from Dialogflow to firebase, and Dialogflow isn't natively used in a sort of questionnaire form. This was important to us because we wanted our users to know that they're more than just a form that they have to fill out with some numbers, their futures really matter to us and that's what we wanted to show. On another end, we spent about 3-4 hours trying to just setup the Apache Tomcat web server that we were planning on using to host the entire Spring MVC. Once we were able to properly get the web server to start running, setting up the MVC was a whole other issue. We had to read a ton of documentation on how to set it up but even then, the sample files that we used to establish the servers had specific formatting which was used within the tomcat server. In the end, trying to learn the whole Spring MVC architecture proved to be too big of an undertaking.

Accomplishments that we're proud of

Properly training the Dialogflow bot and implementing firebase with Dialogflow was definitely an accomplishment that we were proud of. In addition, setting up the tomcat apache server proved to be alot more difficult than we originally thought but learning how to set that up was a huge learning experience we were grateful for. Overall, we're proud of the entire project as all of our group members went from "zero to hero" in at least one part of the project.

What we learned

We learned the do's and donts of web hosting and how to choose the frameworks we want to implement very carefully. We also learned alot about the different API's that were offered by google and other companies as we went through many of them in order to decide which would best fit our project's purpose.

What's next for

We definitely want to fully integrate some of our tools together and create a better experience. Also, with the experience we gained at this hackathon, I think we're all ready to expand on the criteria which can be obtained from Dialogflow. Finally, we want to create a better front end which will display the loan information in a much more aesthetically pleasing way.

One of our teammates may not be able to join the devpost for the hackathon and therefore I can't add her email but I'll stick it to the bottom of this: -- Carmen Wu


Built With

Share this project: