Project Summary

Inspiration

The OECD Better Life Index is a rich but complex dataset that can be difficult to grasp from static reports and simple comparisons generated automatically by Plotly Studio. My inspiration was to create a dynamic, interactive tool that would empower anyone to explore the nuances of well-being using the defining features in the dataset: country, year, age, sex, and education subgroups.

What it does

Wellbeing Compass transforms the dense OECD data into an intuitive dashboard. It allows users to filter data by country, year, and demographics, and then visualize trends, compare population subgroups, and uncover deep patterns using advanced analytics like Probabilistic PCA and correlation heatmaps.

How we built it

I used an iterative, conversational workflow with an AI assistant (Gemini) and Plotly Studio. The process involved directing the AI to perform data cleaning, conduct advanced statistical analyses in R, and design a series of visualizations, progressively building a complete analysis script that serves as the blueprint for the app. Then the R code and insights gained from these analyses were used to structure the prompt for Plotly Studio to generate the app.

Challenges we ran into

The primary challenge was the significant amount of missing data, which required moving beyond standard methods to specialized techniques like Probabilistic PCA. It was also a challenge to craft precise prompts to guide the AI to perform these specific, multi-step statistical analyses correctly and to debug its output.

Accomplishments that we're proud of

We are proud of successfully applying advanced, validated statistical methods to a messy, real-world dataset to extract meaningful insights. We effectively navigated a complex data cleaning and analysis process, resulting in a clear and powerful set of visualizations that tell a compelling story about well-being in OECD countries.

What we learned

This project confirmed that AI is a powerful force multiplier for data analysis, but human expertise is the critical component. We learned that domain knowledge is essential for asking the right questions, and statistical literacy is crucial for guiding the AI and validating its results to ensure the final analysis is both accurate and insightful.

What's next for Well-being Compass

The next step is to develop the R analysis script into a fully deployed, interactive web application using a framework like R Shiny or Plotly Dash. Future plans also include adding more user-driven features, such as allowing users to weight different indicators to create their own custom well-being index and building linear models or casual analysis between correlated variables to investigate the correlated measures.

Built With

• python