-
Home page
-
Form Field
-
Filled form, pop up confirming answer
-
Invalid option pop up for a category question
-
invalid option for digit answer only question
-
Pop up returning AI's salary prediction
-
input box turns white with hover
-
submit button turns white with hover
-
navigation bar links turn bold with hover
Inspiration
As I was doing research for a problem, I found the gender pay gap. Which came as a surprise that first this existed since you would think by now we are passed that, and second that there had been no reliable website existing already to combat it. As I continued my research I found there is bias in data online, a significant pay gap, pay walls blocking access to these predictions that should be accessible to all, and websites not personalized enough to be accurate. Maybe I am selfish, but when I am older I would like to be paid fairly if I am working as much as my male counterparts, and not only for me but for all women, some who don't even know this exists.
What it does
The user gives her degree, hours working daily, field, job level, company size, years of experience, company size and country, the AI will then return her well-deserved salary.
How we built it
Using VS CODE, the backend code consisted of the AI, coded using a Linear Regression Algorithm on PyTorch. The model was then stored into a file for the Flask app to process. The front-end code was built with HTML, CSS and JavaScript, which took the user's inputs, sent it to the Flask app, which encoded the answers, put them through the AI and then sent back the predictions for the user to view. Throughout the coding time, I tried to update my code on Github for version control, showing my code change overtime as I fixed bugs and added to the code.
Challenges we ran into
A major issue my code ran into was that my prediction was returning as undefined, the bug burned at least 2 hours of my time. In the end it turned out to have a few causes some being; my Flask app was not running from a misspelling when defining the app, I was returning a prediction format from Flask that was unreadable to my Frontend, my AI was saved in the wrong file type, and more. Another challenge was the time constraint as running my AI would take at least 5 minutes.
Accomplishments that we're proud of
I am really proud of my whole front end as my strengths lie more in Python and AI. Flask was also pretty new to me and I kept getting errors as I forgot to run my app before testing.
What we learned
In preparation for this hackathon I learned, Flask, Pytorch, JavaScript, HTML, and CSS. I also learned a few tips from the Interviewing Workshop which I really enjoyed, and hope to implement for interviews in the future.
What's next for FairShare
I hope to add a OpenAI for its web scraping so that the users could talk with it, and it could recommend job offers where the women would be paid higher. Eventually, I would hope to make my own full Large Language Model(LLM) that would do the same. Finally, I would add to my Frontend to increase accessibility using tooltips, and replacing alerts with page pop-ups.
Log in or sign up for Devpost to join the conversation.