Inspiration
A company's success depends on employees. We were inspired by ADP's devoted team, seeing the positive impact of their technological advancements. As a team, we are walking towards a career in technology at a company that values our sentiments and values. Thus, we developed the AI Employee Sentiment Analysis which will enable companies to analyze employee sentiments, learn company strengths, tackle their concerns, and proactively consider solutions. We were encouraged by the hope of improving the work environment and company culture as a whole.
What it does
Our project processes employee communications (emails, chats, and feedback) and provides insights on morale trends, burnout signs, and common workplace concerns. Users can select predefined prompts or enter custom ones to generate AI-powered analysis.
How we built it
For this project, we built on what we’ve learned from classwork and past projects, especially when it comes to working with APIs and handling data. We used the Google Gemini AI API to analyze employee feedback, helping us uncover key insights. Since we didn't have pre-existing data, we generated it using the API. Python and Pandas made it easy to process and structure the CSV data, while Streamlit allowed us to create an interactive platform where users could input prompts and see real-time results. To make the front-end visually appealing, we styled it with CSS. One of our main priorities was ensuring smooth integration between the backend and frontend so that everything worked seamlessly from data processing to user interaction.
Challenges we ran into
Learning to use Streamlit: Since we were unfamiliar with it, setting up a functional UI took some trial and error.
Frontend-Backend Connection: Making sure our interface properly communicates with the AI model.
Finding the Right Prompts: Experimenting with various questions to ensure the AI provided meaningful insights.
Finding the Right Color Palette: We wanted our interface to feel professional, inclusive, and easy to read. However, selecting the right combination of colors took time, as some choices reduced readability, while others felt too generic.
Accomplishments that we're proud of
We're really proud of how we brought this AI-powered tool to life, turning raw employee feedback into meaningful insights. Not only did we successfully integrate AI, but we also built an intuitive and visually appealing interface to make it easy to use. We tackled challenges with AI prompts to refine the accuracy of our analysis, ensuring more valuable results. But what we’re most proud of is creating a tool that can genuinely help companies understand employee sentiment, improve workplace satisfaction, and make better HR decisions.
What we learned
We learned the importance of writing precise prompts to get the results we want using AI insights. We also learned how important it is to have properly formatted databases and how to implement AI with front-end development. Learning to work as a team gave us insights on how to troubleshoot coding issues and improve our design.
What's next for this project?
We want to expand our project to incorporate graphs such as pie charts or bar graphs to show visual representations of the analytic reports. Adding real-time tracking to let companies determine company morale dynamically instead of static reports is also something we would like to add in the future. Lastly, implementing a multi-language feature would benefit companies all over the world.
Log in or sign up for Devpost to join the conversation.