🏨 Talk Hotel Data to Me: Date Flexible Viz with AI Assisted Data Queries 🚀 Google Gemini 2.0 (LLM) SQLite LangChain Altair
Problem:
Manual Logs: Management and business operations relied on static, manual logs and spreadsheets.
Visibility Gap: Difficulty spotting long-term trends, comparing seasonal occupancy, or querying specific metrics without extensive manual data digging.
Solution:
Data Engineering: Migrated static logs into a structured SQLite database for high-speed querying.
Natural Language AI: Integrated Google Gemini via LangChain to build a Natural Language-to-SQL engine enabling plain-English queries (e.g., "What was our average rate last weekend?").
Visual Dashboard: Deployed a, cloud-hosted Streamlit application featuring dynamic Altair charts for Occupancy, Sold-Out Pressure, and Maintenance Impact.
Web Application Preview
Short video demonstrating the web application's functionality
Follow Streamlit Link to Web App
⛳ The Golfable Oracle: Personalized Playability Forecasts 🚀 Pandas Plotly
Problem:
Weather Uncertainty: Golfing conditions fluctuate hourly (e.g., pleasant now, rain in 2 hours at the 10th tee), making it difficult to optimize play windows without granular forecasts.
Solution:
Interactive App: Developed a dynamic Streamlit application enabling users to input custom location and playability parameters (min temp, max wind, max precipitation chance).
API Integration: Leveraged the Tomorrow.io API to compute metric differentials, highlight comfortable playing hours on interactive Plotly graphs, and project multi-day golfable hour totals.
Web Application Preview
Short video demonstrating the web application's functionality
Follow Streamlit Link to Web App
💊 Alcohol Treatment: Study Design Explorer with Statistical Insights 🚀 Pandas Matplotlib Numpy Scipy
Problem:
Communication Gap: Scientific studies involve complex design schemas, raw data, and statistical analyses that are often difficult to synthesize and share clearly with broader audiences.
Solution:
Interactive Dashboard: Built an intuitive Streamlit web application breaking down the study design via a visual schematic.
Guided Analysis: Stepped users through group comparisons, statistical tests, and color-coded group manipulations, culminating in interactive, publication-ready key findings visualizations.
Web Application Preview
Short video demonstrating the web application's functionality
Follow Streamlit Link to Web App
📈 The Stock Stack Up: Visualizing Risk and Return Across Stocks 🚀 Yfinance Numpy Pandas Plotly
Problem:
Comparison Limitation: Standard finance platforms isolate individual stocks or fail to benchmark multiple assets effectively on identical timelines, often obscuring risk-vs-return variability.
Solution:
Multi-Asset App: Created a dynamic Streamlit application allowing users to select up to ten tickers and custom timeframes.
Comparative Analytics: Programmed multi-metric Plotly visualizations tracking price changes, percentage shifts, and risk-versus-return dispersion to evaluate historical performance.
Web Application Preview
Short video demonstrating the web application's functionality
Follow Streamlit Link to Web App
🦾 Of Reps & Results: Visualizing Gym Activity Data 🚀 Pandas Numpy Plotly
Problem:
Dataset Interpretation: Public fitness datasets are increasingly accessible, but raw data requires systematic cleaning and contextual interpretation to yield meaningful consumer insights.
Solution:
EDA Dashboard: Developed an interactive Streamlit dashboard mapping out the complete exploratory data analysis workflow, data cleaning steps, and statistical justifications for gym member habits.
Web Application Preview
Short video demonstrating the web application's functionality
Follow Streamlit Link to Web App
🚴 From Rows to Results: Visualizing Big Data with Streamlit 🚀 Pandas Plotly
Problem:
Big Data Complexity: Large-scale analysis (over 144k rows and multiple CSV files) requires robust SQL querying, but raw backend queries lack intuitive visual accessibility for general audiences.
Solution:
SQL + Streamlit Pipeline: Processed multi-file datasets using BigQuery/SQL to structure core metrics, then built an interactive Streamlit dashboard translating large-scale cycling analytics into intuitive visual insights.
Web Application Preview
Follow Streamlit Link to Web App