Landing page
Sign in page
Recording page
Rizz report

Rizzerator

The confidence model that you'll rizzspect
Explore the docs »

Table of Contents

About The Project
- Built With
Getting Started
- Prerequisites
Usage
License
Contact
Acknowledgments

About The Project

Screenshot of Rizzerator landing

The Rizzerator is a machine learning classification model designed to help improve your speech. Using advanced machine learning techniques it provides a confidence score to the inputed audio file, that being how much rizz was detected. The model was trained on hundreds of thousands of speeches and has found nuances in word choice, candence, and emotions to accurately rate the confidence (rizz) of an audio file. The Rizzerator will take you beyond your potential, allowing you to analyze and improve your speech has never been easier - with rizzz.

About our model:

Bank customer service caller audio and agent audio, along with transcipt, used to train model
We used Whisper (OpenAI Library) to transcribe the audio to text
Linguistic feature extractors used: Word2Vec, Glove, BERT
Acoustic feature extractors used: pitch, zero-crossing rate, energy, centroid, spectral spread, spectral entropy

All data used is public access.

(back to top)

Built With

(back to top)

Getting Started

Use the link or www.rizzerator.io clone the repo to get started

Prerequisites

This is an example of how to list things you need to use the software and how to install them (when cloning).

npm sh npm install
go sh go get

(back to top)

Usage

We believe that everyone can improve with the Rizzerator, allowing people to analyze the see the easy corrections to make would allow those who can improve, to improve.

Some use cases:

You have a presentation and would like to do a test run, The Rizzerator wouuld generate a confidence score and suggest possible changes to improve said score based on the input audio.
Job interview coming up soon and want to practice? Run some small talk topics over with the Rizzerator and maximize your chances of employment
Nervous to talk to someone you like? Confess your feelings to them first with the Rizzerator, and boom you're in

(back to top)

License

Distributed under the MIT License. See LICENSE.txt for more information.

(back to top)

Our team

Mauricio Del Castillo - mauriciodelcas30@gmail.com github Hadi Baajour - hadimbaajour@gmail.com github Skyler Estavillo - cryus.estavillo@gmail.com github Tai Tran - ttai8501@gmail.com github

Project Link: https://github.com/JJX30/Rizzerator-App

(back to top)

Acknowledgments

Use this space to list resources you find helpful and would like to give credit to. I've included a few of my favorites to kick things off!

(back to top)

Built With

Submitted to

HackaBull 2023

Created by

Create the front end of the site to communicate with the golang backend. Used styled components, react icons, fetch api. Helped with the fetch of the audio file to the backend, as well as the integration of our python model into the go backend

Mauricio Del Castillo
I designed the logo. I also worked along with Cyrus to help implement the tools that could pick up audio and transcribe it. We worked together to build a model that was able to send info on someone's confidence based off that recording.

taiphlosion
First, I scoured the web for a publicly available dataset that would allow us to have reasonably balanced samples of both confident and unconfident speech. Using Python and a number of its associated libraries, I analyzed audio and transcript file from the Github repository, extracted linguistic and acoustic features, and output these values to a .csv file which could be analyzed by classification models to make future predictions regarding the confidence of speech.

cyrus-estavillo Estavillo
Worked on setting up the backend in Go. Used Gorilla Mux to implement request routers and dispatchers for matching incoming requests to their respective handler. Also used Gorm to initialize migration and creation of the database schema.

Private user