Airlines Delay Prediction using DL

[ananyag309]

Deep Learning Simplified Repository (Proposing new issue)

🔴 Project Title : Airlines Delay Prediction

🔴 Aim : Airlines Delay Prediction is a machine learning project aimed at predicting flight delays. By predicting potential delays, airlines can manage their schedules better and passengers can plan their travels more efficiently.

🔴 Dataset : It includes historical flight data with features such as departure time, arrival time, weather conditions, and other relevant factors.

🔴 Approach : The project uses several machine learning models to predict flight delays, including:

Linear Regression
Decision Trees
Random Forest
Gradient Boosting
Neural Networks

---

📍 Follow the Guidelines to Contribute in the Project :

• You need to create a separate folder named as the Project Title.
• Inside that folder, there will be four main components.
  • Images - To store the required images.
  • Dataset - To store the dataset or, information/source about the dataset.
  • Model - To store the machine learning model you've created using the dataset.
  • requirements.txt - This file will contain the required packages/libraries to run the project in other machines.
• Inside the Model folder, the README.md file must be filled up properly, with proper visualizations and conclusions.

---

🔴🟡 Points to Note :

• The issues will be assigned on a first come first serve basis, 1 Issue == 1 PR.
• "Issue Title" and "PR Title should be the same. Include issue number along with it.
• Follow Contributing Guidelines & Code of Conduct before start Contributing.

---

✅ To be Mentioned while taking the issue :

• Full name : Ananya Gupta
• GitHub Profile Link : https://github.com/ananyag309
• Approach for this Project :
• What is your participant role? (Mention the Open Source program) GSSOC

---

Happy Contributing 🚀

All the best. Enjoy your open source journey ahead. 😎

[github-actions]

Thank you for creating this issue! We'll look into it as soon as possible. Your contributions are highly appreciated! 😊

[abhisheks008]

Machine learning models will not here, can you update your approach with some deep learning architectures and come up again.
@ananyag309

[Likhith2612]

Hi @ananyag309 , I have tried the project you can see it once .
for linear regression, the results are quite low.
have tried svm and gradient boost.
code.pdf
Airlines_Dataset.csv

[abhisheks008]

Hi @ananyag309 , I have tried the project you can see it once . for linear regression, the results are quite low. have tried svm and gradient boost. code.pdf Airlines_Dataset.csv

This repository mainly focuses on deep learning models, hence you need to go with the deep learning models not with the machine learning ones.

[Likhith2612]

ok.

[SibhiSS]

Full name: Sibhi S
GitHub Profile: https://github.com/SibhiSS
Approach: Will implement a comparative study using Random Forest and Gradient Boosting as baselines alongside a Neural Network.
Participant role: GSSOC

[abhisheks008]

Full name: Sibhi S GitHub Profile: https://github.com/SibhiSS Approach: Will implement a comparative study using Random Forest and Gradient Boosting as baselines alongside a Neural Network. Participant role: GSSOC

Please share your approach based on the deep learning methods and re-consider the same.

[Rudrani-Mukherjee]

Hi @abhisheks008, I listed my approach down, please go through it
and assign me under tag GSSoC 2026.

Full name: Rudrani Mukherjee
GitHub Profile Link: https://github.com/Rudrani-Mukherjee
Email ID: rudranimukherjee063@gmail.com
Participant ID: fe7e06e6-8b3b-4cf1-b467-3d298c3ee657

Approach for this Project: I will begin with data loading and comprehensive EDA — analyzing flight patterns, delay distributions, busiest routes, peak delay hours, and weather impact using Matplotlib and Seaborn (histograms, heatmaps, box plots, correlation matrix). After preprocessing (handling missing values, encoding categorical features like airline/airport codes, feature scaling and engineering delay-related features), I will implement and compare these models:

1. ANN — primary deep learning model for delay prediction
2. Random Forest — to capture non-linear feature relationships
3. Gradient Boosting (XGBoost) — for high accuracy prediction
4. LSTM — to capture temporal/sequential patterns in flight data
   Each model will be evaluated using RMSE, MAE, accuracy score, and loss curves. A final comparison table will highlight the best performing model.
   Participant Role: Contributor, GSSoC 2026

Looking forward to contributing!

[abhisheks008]

Hi @abhisheks008, I listed my approach down, please go through it and assign me under tag GSSoC 2026.

Full name: Rudrani Mukherjee GitHub Profile Link: https://github.com/Rudrani-Mukherjee Email ID: rudranimukherjee063@gmail.com Participant ID: fe7e06e6-8b3b-4cf1-b467-3d298c3ee657

Approach for this Project: I will begin with data loading and comprehensive EDA — analyzing flight patterns, delay distributions, busiest routes, peak delay hours, and weather impact using Matplotlib and Seaborn (histograms, heatmaps, box plots, correlation matrix). After preprocessing (handling missing values, encoding categorical features like airline/airport codes, feature scaling and engineering delay-related features), I will implement and compare these models:

1. ANN — primary deep learning model for delay prediction
2. Random Forest — to capture non-linear feature relationships
3. Gradient Boosting (XGBoost) — for high accuracy prediction
4. LSTM — to capture temporal/sequential patterns in flight data
   Each model will be evaluated using RMSE, MAE, accuracy score, and loss curves. A final comparison table will highlight the best performing model.
   Participant Role: Contributor, GSSoC 2026

Looking forward to contributing!

Can you suggest deep learning models rather than machine learning techniques.

[Rudrani-Mukherjee]

Hi @abhisheks008, I listed my approach down, please go through it and assign me under tag GSSoC 2026.
Full name: Rudrani Mukherjee GitHub Profile Link: https://github.com/Rudrani-Mukherjee Email ID: rudranimukherjee063@gmail.com Participant ID: fe7e06e6-8b3b-4cf1-b467-3d298c3ee657
Approach for this Project: I will begin with data loading and comprehensive EDA — analyzing flight patterns, delay distributions, busiest routes, peak delay hours, and weather impact using Matplotlib and Seaborn (histograms, heatmaps, box plots, correlation matrix). After preprocessing (handling missing values, encoding categorical features like airline/airport codes, feature scaling and engineering delay-related features), I will implement and compare these models:

1. ANN — primary deep learning model for delay prediction
2. Random Forest — to capture non-linear feature relationships
3. Gradient Boosting (XGBoost) — for high accuracy prediction
4. LSTM — to capture temporal/sequential patterns in flight data
   Each model will be evaluated using RMSE, MAE, accuracy score, and loss curves. A final comparison table will highlight the best performing model.
   Participant Role: Contributor, GSSoC 2026

Looking forward to contributing!

Can you suggest deep learning models rather than machine learning techniques.

Thank you for the feedback!

I will replace ML models with the following Deep Learning approaches for Airlines Delay Prediction:

1. ANN (Artificial Neural Network) — baseline DL model for tabular flight data, capturing complex non-linear relationships between features like departure time, weather, and route
2. LSTM (Long Short-Term Memory) — ideal for capturing sequential/temporal patterns in historical flight delay data over time
3. GRU (Gated Recurrent Unit) — efficient alternative to LSTM with fewer parameters, faster training while maintaining similar accuracy for time-series patterns
4. 1D-CNN (1D Convolutional Neural Network) — extracts local feature patterns from flight sequences, works surprisingly well on structured tabular data
   Each model will be evaluated using RMSE, MAE, and loss curves with a final comparison table highlighting the best performing model. This combination ensures both temporal patterns and complex feature relationships are captured effectively for the best prediction accuracy.

Looking forward to contributing!

[SibhiSS]

Full name: Sibhi S GitHub Profile: https://github.com/SibhiSS Approach: Will implement a comparative study using Random Forest and Gradient Boosting as baselines alongside a Neural Network. Participant role: GSSOC

Please share your approach based on the deep learning methods and re-consider the same.

Sure! Revised approach focusing on DL,
Will use the BTS airline delay dataset from Kaggle. EDA covers delay distributions, busy routes, and time/weather correlations.

For models, FNN as the DL baseline, 1D CNN to capture local feature patterns, LSTM for temporal dependencies across flight schedules, and a TabTransformer since it handles mixed tabular features really well. Random Forest and Gradient Boosting stay as classical baselines for comparison.

Evaluation will include RMSE, MAE, R² and if framed as classification — accuracy, F1, and confusion matrices.

[abhisheks008]

Hi @abhisheks008, I listed my approach down, please go through it and assign me under tag GSSoC 2026.
Full name: Rudrani Mukherjee GitHub Profile Link: https://github.com/Rudrani-Mukherjee Email ID: rudranimukherjee063@gmail.com Participant ID: fe7e06e6-8b3b-4cf1-b467-3d298c3ee657
Approach for this Project: I will begin with data loading and comprehensive EDA — analyzing flight patterns, delay distributions, busiest routes, peak delay hours, and weather impact using Matplotlib and Seaborn (histograms, heatmaps, box plots, correlation matrix). After preprocessing (handling missing values, encoding categorical features like airline/airport codes, feature scaling and engineering delay-related features), I will implement and compare these models:

1. ANN — primary deep learning model for delay prediction
2. Random Forest — to capture non-linear feature relationships
3. Gradient Boosting (XGBoost) — for high accuracy prediction
4. LSTM — to capture temporal/sequential patterns in flight data
   Each model will be evaluated using RMSE, MAE, accuracy score, and loss curves. A final comparison table will highlight the best performing model.
   Participant Role: Contributor, GSSoC 2026

Looking forward to contributing!

Can you suggest deep learning models rather than machine learning techniques.

Thank you for the feedback!

I will replace ML models with the following Deep Learning approaches for Airlines Delay Prediction:

1. ANN (Artificial Neural Network) — baseline DL model for tabular flight data, capturing complex non-linear relationships between features like departure time, weather, and route
2. LSTM (Long Short-Term Memory) — ideal for capturing sequential/temporal patterns in historical flight delay data over time
3. GRU (Gated Recurrent Unit) — efficient alternative to LSTM with fewer parameters, faster training while maintaining similar accuracy for time-series patterns
4. 1D-CNN (1D Convolutional Neural Network) — extracts local feature patterns from flight sequences, works surprisingly well on structured tabular data
   Each model will be evaluated using RMSE, MAE, and loss curves with a final comparison table highlighting the best performing model. This combination ensures both temporal patterns and complex feature relationships are captured effectively for the best prediction accuracy.

Looking forward to contributing!

Start working on it @Rudrani-Mukherjee. Assigned to you!