VizWiz | Devpost

Inspiration

We were inspired by the need to empower businesses to optimize resource usage and improve sustainability. Property managers often struggle with making informed decisions from complex data sets. Our goal was to simplify their decision-making process by combining visual insights with an easy-to-use voice assistant. This tool enables them to easily track and improve sustainability in their office spaces.

What it does

Turning data chaos into clarity with vibrant graphs and a friendly voice chatbot ready to talk about your stats. The tool provides detailed visualizations of inputted files to give clear insights. Property managers can interact with the voice assistant to get real-time suggestions and improvements for their office spaces. The assistant also helps users create more sustainable and efficient office environments through continuous conversation.

How we built it

We built the project with React.js for the frontend, allowing users to upload their data files easily. These files are stored on Piñata, a decentralized file storage system, for efficient and secure data handling. The backend is powered by Flask and Python, which processes the data and interfaces with OpenAI's assistant models to create a virtual property manager. Google Chart API is used to generate dynamic visualizations based on the files uploaded by users, with all components integrated seamlessly into the React application using Vite.

Challenges we ran into

Integrating Piñata for decentralized file storage while ensuring fast access to large datasets required careful optimization. Managing the seamless data flow between the frontend, backend, and the LLM posed challenges, especially with CORS and real-time API calls.

Accomplishments that we're proud of

We successfully implemented decentralized file storage with Piñata for secure, scalable data handling. The integration of advanced visualizations and a conversational LLM assistant delivers actionable insights in an intuitive interface.

What we learned

We deepened our understanding of decentralized file storage with Pinata, managing large datasets for real-time analysis and combining LLMs with user-friendly interfaces. This project reinforced our ability to build scalable, cloud-native applications.

What's next for VizWiz

Future plans include integrating predictive analytics, expanding support for more data types, and refining the LLM to provide industry-specific insights. We also aim to enhance real-time data pipelines and explore additional blockchain-based solutions for secure data management.