Inspiration
Have you ever been scared of messing up while presenting in front of your peers. Terrified your boss will fire you if you miss one more key points in front of shareholders? Well, if you do, then Speech Savvy can solve all your problems! No more crying in front of your parents, no more shying away from speaking in public. Use Speech Savvy, and you will be a better speaker, guaranteed.
What it does
Our software takes the text file of your speech, what you want to say, and compares it against the audio file of what you said when you practiced. The audio file is converted to text, and using our complex algorithm, we compare the files and output percent accuracy of the two speeches, percent and number of keywords hit, number of and list of keywords missed, the amount of filler words used, and improvement from previous attempt.
How we built it
We used amazon transcribe to convert the audio file into a text file. Then, we compared the inputted text file with the converted one and output the statistics. We used python as our coding platform, because it has a nltk library which allowed us to use tokenizer to split the text into an array of individual words and then used it to filter out stop words, such as "uh", "um" etc. amazon web services s3 and transcribe
Challenges we ran into
At first our team was conflicted about what platform to use to create our website and to convert the audio into text. After extensive research, trial and error, and help from our mentors we unanimously decided upon using AWS software as it was the most user friendly option. We then ran into the issue of getting the website to record audio into an mp3 file, as well as using the information in the website in our python program.
Accomplishments that we're proud of
We are proud of creating an attractive user-friendly website which can easily perform the above tasks and display the output in an aesthetically pleasing way.
What we learned
We learned python and libraries and how to use amazon transcribe to convert audio files into text. Also, we learned how to create a website using aws.
What's next for SpeechSavvy
In the future, we will implement methods to output more helpful statistics. We also hope to implement provide feedback on many more features including volume of speaker, speed of speaking, and tone. Additionally, another feature we will implement is the ability to recognize multiple speakers in one audio recording.
Log in or sign up for Devpost to join the conversation.