Project Title and Date:
Flight Delays Analysis
Sep 2024 – Present
Project Overview
This project analyzed over 170,000 U.S. flight records to identify patterns in delays, uncover root causes, and propose actionable solutions to improve the experience of the average flier. By narrowing the dataset to the top six airlines (Delta, American, Southwest, United, Alaska, and JetBlue) and the 50 busiest airports, the analysis focused on the travel behavior of everyday passengers. Special emphasis was placed on the holiday season and winter months, particularly December, which saw the highest delays.
Key Features & Methodology
Exploratory Data Analysis (EDA)
Analyzed trends in delays across years, airlines, and airports to uncover seasonal and carrier-specific patterns.
Visualized insights using Python libraries (Matplotlib, Seaborn) to identify Southwest Airlines as the leader in delayed departures among budget carriers.
Focused on winter and holiday delays, revealing December as the most delay-prone month.
Predictive Modeling & Regression Analysis
Built Linear Regression, Decision Tree, and Random Forest models to predict delays based on factors such as weather, carrier, airport, time of year, and delay types.
Leveraged the models to understand key predictors of delays and assess the trade-offs between cost and reliability for different airlines.
Fine-tuned models for accuracy and interpretability, providing actionable insights for the airline industry.
Clustering and Geospatial Analysis
Applied K-means clustering to categorize airports based on delay likelihood, identifying operational weaknesses at high-traffic hubs.
Incorporated geospatial techniques to map delay hotspots and prioritize areas for intervention.
Winter Delay Diagnosis
Focused on the impact of delays during the holiday season, addressing common travel scenarios such as family reunions and winter vacations.
Explored delay types, uncovering that most delays were attributed to carrier and late aircraft issues.
Comprehensive Report
Authored a 15-page report summarizing findings and strategic recommendations tailored for airlines and airports.
Included visualizations, predictive model results, and actionable steps to reduce delays and improve customer satisfaction.
Technical Skills
Programming: Python, Pandas, Scikit-learn, Matplotlib, Seaborn
Machine Learning: Linear Regression, Decision Trees, Random Forest, K-means clustering
Data Analysis: Exploratory Data Analysis (EDA), correlation analysis, geospatial mapping
Reflection
This project provided a deep understanding of the factors influencing flight delays, particularly during the busiest times of the year. By integrating machine learning models with exploratory and geospatial analysis, the project offered actionable insights for improving reliability and efficiency. These findings lay the groundwork for enhancing the flying experience for all passengers, especially during peak travel seasons, emphasizing the importance of data-driven strategies and collaboration across the airline industry.