Inspiration
Do you struggle with understanding other people and recognising social cues? Have you ever been nervous going to an event where you don't know anyone? Or tried to join a conversation where everyone knows each other and it's so cliquey, you can't join in? Our project aims to tackle that. We want to build something that helps you come out of social events feeling confidence because you've started your own legacy.
What it does
Our widget aims to sit on the user's desktop as a social helper. When clicked, the window opens up with three options to choose from:
- Discuss - from either a transcript or uploaded recording of an interaction, analyse a prior conversation and discuss nuances and possible improvements
- Simulate - talk to the bot about what you want to achieve before different AI personalities appear acting in an excluding manner within a practice scenario in which you will deescalate and elevate the conversation
- Progress - see previous chats and after enough hours, see how the bot expects your next IRL interaction will go
How we built it
Using Python as our primary language, we made use of Google Gemini’s AI models and their Natural Language Understanding capabilities to take into account conversational context and recognise and analyse conversational tone, emotional sentiment and personal intentions of speakers within a conversation. We used PyQt6 to build a GUI. Sprites and buttons were hand drawn by members.
Challenges we ran into
We found time management a bit of a challenge as there were so many challenges and events we wanted to participate in, making us nervous about the deadline. We overcame this by having clear communication, both verbally and by creating a discord group. This also allowed us to easily share files for efficiency.
Accomplishments that we're proud of
Working with new tools and creating a functioning application with no errors in our tight time constraint is a massive accomplishment to us. Alongside all the other events, challanges and creating meaningful friendships in the process.
What we learned
We have learned new tools such as Gemini AI, the PyQt library for the GUI, Figma for wireframes and much more. As a team, we definetly worked with at least one new tool we can apply to future projects.
What's next for Clique Into Place
Due to our strict time constraints, there were a few features we were not able to implement which we would love to for future develpement. One of which being voice generating AI that can help with our simulation, simulating a real life conversations, using a range of tone and detecting patterns of sarcasm for example.
Log in or sign up for Devpost to join the conversation.