DataWhisper

• 

  DataWhisper landing page

Inspiration

We wanted to make Excel as effortless as having a conversation. Many students and educators struggle with formulas, data analysis, and visualization, so we envisioned a system where you could simply speak or type natural language commands and the spreadsheet would perform the operations automatically. This inspired the creation of DataWhisper, an AI-powered spreadsheet assistant that makes working with student and academic data more intuitive and intelligent.

What it does

DataWhisper allows users to perform spreadsheet operations through natural language prompts instead of manual formulas or menu navigation. Users can give instructions such as:

“Add a new column for average marks”

“Delete the third row”

“Combine first and last names”

“Return the students who got the highest marks”

“List the students who scored above average”

“Show me the top 5 performers in the class”

“Display a chart of class performance over time”

“Find the student with the lowest attendance”

The system understands these commands in context and executes them automatically. It can also analyze student performance data, generate visualizations, and provide insights using the Gemini API. For hands-free interaction, DataWhisper supports voice input and output through the ElevenLabs API, allowing users to interact with their data naturally.

How we built it

We built DataWhisper using a modular architecture:

Frontend: A user interface built with React to handle text or voice input and display updated spreadsheet results.

Backend: Flask framework integrated with the Gemini API to interpret natural language and convert it into structured spreadsheet operations.

Voice Interaction: Integrated ElevenLabs API to process voice inputs and deliver voice-based responses, enabling conversational interaction.

Visualization Layer: Uses Gemini and Chart.js to automatically generate charts, graphs, and summaries from student data.

Challenges we ran into

Translating vague or multi-step natural language instructions into precise spreadsheet actions.

Handling complex queries involving ranking, averages, or filtering without errors.

Generating accurate and meaningful charts that correctly represent academic data.

Managing multiple API integrations efficiently within limited time.

Ensuring data consistency during live updates and automation flows.

Accomplishments that we're proud of

Built a functional, intelligent spreadsheet assistant capable of understanding academic and performance-related prompts.

Achieved reliable prompt-to-action execution across a wide range of student data use cases.

Integrated natural language understanding, data visualization, and voice interaction within one cohesive platform.

Completed end-to-end development, testing, and workflow automation successfully within the hackathon timeframe.

What we learned

How to combine large language models with structured data systems for logical and analytical tasks.

The importance of effective prompt design for handling educational data queries accurately.

How conversational interfaces can make complex data operations more accessible to students and educators.

What's next for DataWhisper

Expanding support for platforms like Google Sheets and Airtable to reach a wider audience.

Adding smart grading analytics, student progress tracking, and predictive performance insights.

Enabling real-time collaboration so teachers and students can interact with the same dataset using natural language.

Introducing a lightweight browser extension for quick integration with classroom spreadsheets.

Exploring adaptive learning analytics that can summarize student strengths and weaknesses automatically.

Built With

• eleven-labs
• flask
• gemini
• javascript
• python
• react