Dataset Overview

Dataset Name

• Drive Me Crazy

Technical Details

• Submitted Files: dataset_analysis.ipynb, presentation.txt, lookerstudio_url.txt, drive_me_crazy_tradi.ipynb, drive_me_crazy_pdformer.ipynb
• Programming Language: Determined by the participating Bootcamp, e.g., JavaScript, Ruby, Python, Java, C++, Rust, etc.

Project Objectives

• Develop a traffic flow forecasting system based on the “Propagation Delay‑Aware Dynamic Long‑Range Transformer” method.
• Predict urban traffic flow patterns using the PeMS04, PeMS07, PeMS08, and NYCTaxi datasets.
• Build an interactive dashboard with Looker Studio for visualizing and presenting traffic flow forecasts.

Methodology

• Data Integration: Merge PeMS04, PeMS07, PeMS08, and NYCTaxi datasets.
• Data Preprocessing: Clean, aggregate, and preprocess the data.
• Model Adaptation: Customize the “Propagation Delay‑Aware Dynamic Long‑Range Transformer” model.
• Model Training: Train the model using appropriate data‑science libraries.
• Model Evaluation: Evaluate the model with metrics such as MAE and RMSE.
• Comparative Analysis: Compare the new model with traditional traffic forecasting methods.
• Delay Impact Analysis: Assess how propagation delay affects forecasting accuracy.
• Transformer Efficiency Analysis: Evaluate the effectiveness of the dynamic long‑range transformer in capturing traffic patterns across the various datasets.

Deliverables

• Dataset analysis report.
• Traditional traffic flow forecasting model (Jupyter notebook).
• Implementation of the “Propagation Delay‑Aware Dynamic Long‑Range Transformer” model (Jupyter notebook).
• Report on the impact of propagation delay on forecasting accuracy.
• Interactive dashboard developed with Looker Studio.
• Model performance comparison report.
• Insight report on the efficiency of the dynamic long‑range transformer in capturing urban traffic patterns.