Titanic Survival Analytics

Titanic Survival Analytics Dashboard

Inspiration

I wanted to explore the Titanic disaster using data.
The goal was to find patterns in who survived and build a dashboard that shows insights clearly.

What it does

Shows four main sections:
1. Historical Review (what happened)
2. Root Cause Analysis (why it happened)
3. Scenario Forecasting (what might happen)
Includes interactive charts: histograms, donut charts, violin plots, and heatmaps
Gives easy-to-understand insights from the data

How we built it

Everything was built using Plotly Vibe Code. I provided a single comprehensive AI prompt, and Vibe Code handled:
Loading and cleaning the Titanic dataset
Splitting data into training and test sets
Training three ML models: Random Forest, Decision Tree, Linear SVC
Evaluating models with Accuracy, Precision, and Recall
Building a four-tab interactive dashboard with CYBORG dark theme, KPIs, charts, and prediction inputs

Challenges we ran into

Some charts overlapped, requiring edit to fix in Vibe Code
Missing Strategic Recommendations tab: even though it was in the instructions, Vibe Code did not generate it automatically, so it would need to be added manually ## Accomplishments that we're proud of Fully functional interactive dashboard built entirely in Vibe Code
High-quality visuals with CYBORG dark theme and clear KPIs
Turned messy Titanic data into actionable insights without writing manual code

What we learned

How to trust AI to automate end-to-end workflows
Importance of clean data and proper feature engineering
How to compare multiple ML models and evaluate them effectively
How to present data insights clearly in an interactive dashboard ## What's next for Titanic Survival Analytics
Add the Strategic Recommendations tab manually, with four actionable insights
Include additional ML models and hyperparameter tuning
Add better visualizations for survival probabilities