maorsoul's picture
Update README.md
59eb913 verified
## ๐ŸŽฅ Project Video Walkthrough
<video controls width="720">
<source src="https://huggingface.co/maorsoul/flight-delay-predictor/resolve/main/video1534610661.mp4" type="video/mp4">
Your browser does not support the video tag.
</video>
# โœˆ๏ธ Flight Delay Predictor
## ๐Ÿ“Œ Dataset Overview
For this project, I worked with the **2018 US Flight Delays & Cancellations** dataset.
This dataset contains detailed information about **over 7 million domestic flights in the United States**, including:
* Flight dates and times
* Departure and arrival delays
* Airline carrier codes
* Origin and destination airports
* Distance and air time
* Cancellation and diversion information
* Various time-related features (month, day, day of week, scheduled times, etc.)
To keep the project computationally manageable, I selected a **random sample of 20,000 rows** from the full dataset.
This sample size still preserves meaningful variation in delays, airlines, and airports, allowing for effective modeling without heavy computation.
**Main target variable:**
`ArrDelay` โ€“ the arrival delay in minutes.
This continuous variable was used first for a regression problem, and later converted into classes for a classification task.
**Goal of the project:**
1. Predict arrival delay using regression models.
2. Reframe the problem into classification (high delay vs. low delay).
3. Compare models and deploy the best-performing classifier/regressor to HuggingFace.
The project walks through the full ML process:
* Data loading & cleaning
* EDA
* Feature engineering
* Model training
* Evaluation
* Selecting a winner
* Exporting the model
# ๐Ÿ“Š 2. Exploratory Data Analysis (EDA)
In this section we explored:
* Total rows, columns
* Data types
* Missing values
* Basic statistical patterns
* Target variable behavior before classification
**Main actions performed:**
* Loaded 20,000 rows from the 2018 dataset
* Removed irrelevant fields (like tail IDs)
* Verified missing values and cleaned them
* Verified numerical ranges to detect odd values
* Converted original delay (`ArrDelay`) into the classification target `y_class`
* Split into 80% train, 20% test
โฌ‡๏ธ
![ืฆื™ืœื•ื ืžืกืš 2025-11-29 ื‘-9.37.33](https://cdn-uploads.huggingface.co/production/uploads/69183cd79510e3441ef86afc/H5TkmTdvamGzCX3tnkWbK.png)
![ืฆื™ืœื•ื ืžืกืš 2025-11-29 ื‘-9.38.47](https://cdn-uploads.huggingface.co/production/uploads/69183cd79510e3441ef86afc/ATuj1DhNFu4IOKADBVvfT.png)
![ืฆื™ืœื•ื ืžืกืš 2025-11-29 ื‘-9.39.05](https://cdn-uploads.huggingface.co/production/uploads/69183cd79510e3441ef86afc/uVzJysvmUNKrI6dGyDxJU.png)
![ืฆื™ืœื•ื ืžืกืš 2025-11-29 ื‘-9.39.25](https://cdn-uploads.huggingface.co/production/uploads/69183cd79510e3441ef86afc/dFULLmyeCowD54qkHMV3J.png)
![ืฆื™ืœื•ื ืžืกืš 2025-11-29 ื‘-9.39.41](https://cdn-uploads.huggingface.co/production/uploads/69183cd79510e3441ef86afc/ecNxcebQio2SOgl63r93a.png)
![ืฆื™ืœื•ื ืžืกืš 2025-11-29 ื‘-9.39.57](https://cdn-uploads.huggingface.co/production/uploads/69183cd79510e3441ef86afc/-G9t6hG5_-q9pBHxqN7rT.png)
![ืฆื™ืœื•ื ืžืกืš 2025-11-29 ื‘-9.40.08](https://cdn-uploads.huggingface.co/production/uploads/69183cd79510e3441ef86afc/cL1WcEpM2edSFbPHKSiUo.png)
![ืฆื™ืœื•ื ืžืกืš 2025-11-29 ื‘-9.40.19](https://cdn-uploads.huggingface.co/production/uploads/69183cd79510e3441ef86afc/oY-wiihgzmlMtMvqzIZFK.png)
![ืฆื™ืœื•ื ืžืกืš 2025-11-29 ื‘-9.40.29](https://cdn-uploads.huggingface.co/production/uploads/69183cd79510e3441ef86afc/cYOZt4qv4fOxWr8RxQkfu.png)
### **Insert dataset head or summary as an image**
![ืฆื™ืœื•ื ืžืกืš 2025-11-29 ื‘-9.41.55](https://cdn-uploads.huggingface.co/production/uploads/69183cd79510e3441ef86afc/OzNiiXLyr8accJYlArisL.png)
# ๐Ÿ” 3. Baseline Model
In this phase we studied the patterns behind delay behavior.
### What we analyzed:
* **Distribution of arrival delays**
Helps understand skew, outliers, and how reasonable our classification threshold is.
* **Correlation between numerical features**
Found that distance and scheduled times impact delays but not extremely strongly.
* **Delay behavior by airline**
Some airlines have significantly more variability in delays.
* **Time of day vs delay**
Late-day flights tend to accumulate more delays.
* **Outlier detection using Z-score**
Removed unrealistic delays > ยฑ3 standard deviations.
### Why it matters:
EDA allowed us to understand which features influence delays and how noisy the data is.
This guided feature engineering and reduced overfitting risk.
โฌ‡๏ธ
### **Place graphs here**
![ืฆื™ืœื•ื ืžืกืš 2025-11-29 ื‘-9.44.08](https://cdn-uploads.huggingface.co/production/uploads/69183cd79510e3441ef86afc/M3ETCvLN0Rf_AItthFbw3.png)
![ืฆื™ืœื•ื ืžืกืš 2025-11-29 ื‘-9.44.28](https://cdn-uploads.huggingface.co/production/uploads/69183cd79510e3441ef86afc/-m8eDsCZWlH-AxGvrtZgS.png)
![ืฆื™ืœื•ื ืžืกืš 2025-11-29 ื‘-9.44.41](https://cdn-uploads.huggingface.co/production/uploads/69183cd79510e3441ef86afc/rXCdxRxcnJap6b9U28-d9.png)
# ๐Ÿ› ๏ธ 4. Feature Engineering
Feature engineering was critical for improving model quality.
### Done in this step:
#### **1. One-Hot Encoding for categorical features**
* Airline
* Origin airport
* Destination airport
* Day of Week
* Cancellation field
This expanded the dataset into thousands of columns but preserved categorical meaning.
#### **2. Scaling important numerical fields**
* Distance
* CRSDepTime
* CRSArrTime
* AirTime
Scaling prevents models like Logistic Regression and Gradient Boosting from being biased by large numeric ranges.
#### **3. PCA (optional)**
Used only for visualization; helped validate that the classes are somewhat separable.
#### **4. K-Means clustering (optional exploratory step)**
Cluster labels added as an experimental feature to see if they help models (they had mild impact).
โฌ‡๏ธ
### **Place FE graphs here**
![ืฆื™ืœื•ื ืžืกืš 2025-11-29 ื‘-9.45.11](https://cdn-uploads.huggingface.co/production/uploads/69183cd79510e3441ef86afc/T7pjOhFJL1Zn54OroFK9T.png)
![ืฆื™ืœื•ื ืžืกืš 2025-11-29 ื‘-9.45.26](https://cdn-uploads.huggingface.co/production/uploads/69183cd79510e3441ef86afc/Tq6yVLGH-w8tLty1rbNQG.png)
# ๐Ÿค– 5. Models Trained
We compared **three supervised classification models**:
### โœ” Logistic Regression
* Simple baseline
* Fast, linear, interpretable
* Surprisingly produced perfect predictions (overfitting to clean, thresholded labels)
### โœ” Random Forest Classifier
* Non-linear
* Handles high-dimensional data
* Good but struggled with high-delay recall
### โœ” Gradient Boosting Classifier
* Ensemble of weak learners
* Best real-world performance
* Most balanced precisionโ€“recall
* Strong against noise
* Best generalization to unseen data
โฌ‡๏ธ
### **Insert models summary image**
![ืฆื™ืœื•ื ืžืกืš 2025-11-29 ื‘-9.45.46](https://cdn-uploads.huggingface.co/production/uploads/69183cd79510e3441ef86afc/LWGxTGHU-gYW2QRFOjhm_.png)
![ืฆื™ืœื•ื ืžืกืš 2025-11-29 ื‘-9.46.01](https://cdn-uploads.huggingface.co/production/uploads/69183cd79510e3441ef86afc/WelQ-fbqravyTW1nYv4pW.png)
![ืฆื™ืœื•ื ืžืกืš 2025-11-29 ื‘-9.46.11](https://cdn-uploads.huggingface.co/production/uploads/69183cd79510e3441ef86afc/kA863njS1KJj4ZvvIFvAq.png)
# ๐Ÿ† 6. Winning Model
The selected model is:
# **๐ŸŒŸ Gradient Boosting Classifier**
### Why this one?
* Best tradeoff between false positives and false negatives
* Highest real F1-score
* Handles imbalanced patterns better
* Robust to feature noise and outliers
* Most realistic generalization
## 7. Regression-to-Classification
### 7.1 Creating Classes from the Numeric Target (Median Split)
In this part we reframed the original regression target **ArrDelay** into a
binary classification target.
We computed the **median arrival delay on the training set** (โ‰ˆ โˆ’5 minutes) and
used it as a threshold:
- **Class 0 โ€“ Low delay:** `ArrDelay < median`
(flight is on time or earlier than a typical flight in the dataset).
- **Class 1 โ€“ High delay:** `ArrDelay โ‰ฅ median`
(flight is more delayed than a typical flight).
The same rule was applied to both **train and test** targets, using the **same
engineered features** as in the regression part.
This keeps the classification task aligned with the original question:
> *โ€œHow large will the arrival delay be?โ€*
now phrased as
> *โ€œWill this flight have a higher-than-typical delay or not?โ€*
### 7.2 Checking Class Balance
After creating the classes, we examined their distribution:
- **Training set:**
about **50.6% High delay (Class 1)** and **49.4% Low delay (Class 0)**.
- **Test set:**
about **51.3% Low delay (Class 0)** and **48.7% High delay (Class 1)**.
The classes are therefore **well balanced**, and no class is clearly
under-represented.
Because of this balance, **accuracy** is already informative, but to avoid
being misled in edge cases and to keep the focus on the โ€œHigh delayโ€ class,
we mainly compared models using the **F1-score** (which combines precision and
recall for the positive class).
๐Ÿ‘‰ *Here I will insert a bar plot (or table screenshot) of the class
distribution in train/test.*
![ืฆื™ืœื•ื ืžืกืš 2025-11-29 ื‘-9.55.24](https://cdn-uploads.huggingface.co/production/uploads/69183cd79510e3441ef86afc/fSo9lYbeNOK6_8qrBtFRc.png)
![ืฆื™ืœื•ื ืžืกืš 2025-11-29 ื‘-9.55.39](https://cdn-uploads.huggingface.co/production/uploads/69183cd79510e3441ef86afc/-kh6L76mQaE9tJv4nymxA.png)
## 8. Train & Evaluate Classification Models
### 8.1 Precision vs. Recall โ€” What Matters More?
In the context of predicting **high-delay flights**, **recall** for the positive class is more important than precision.
The reason:
Missing a truly delayed flight (false negative) is operationally worse than mistakenly flagging
an on-time flight as delayed (false positive).
A missed severe delay can lead to missed connections, poor customer experience, and scheduling disruptions,
while a false alarm only causes minor adjustments like extra buffer time.
---
### 8.1 False Positives vs. False Negatives โ€” Which Is Worse?
- A **false positive** means predicting โ€œhigh delayโ€ when the flight is actually low-delay.
- A **false negative** means predicting โ€œlow delayโ€ when the flight is actually highly delayed.
In our task, **false negatives are more critical**, because they leave planners unprepared for major delays.
False positives are less harmful โ€” they may cause unnecessary caution, but do not create operational failures.
---
### 8.2 Training Three Classification Models
We trained and evaluated three different models from scikit-learn, using the same engineered features
and the binary target created in Part 7:
1. **Logistic Regression**
2. **Random Forest Classifier**
3. **Gradient Boosting Classifier**
๐Ÿ‘‰ *Insert model training diagram or screenshots of code here (optional).*
![ืฆื™ืœื•ื ืžืกืš 2025-11-29 ื‘-9.57.44](https://cdn-uploads.huggingface.co/production/uploads/69183cd79510e3441ef86afc/vHzlqE8vnf7tRBxACgY-V.png)
![ืฆื™ืœื•ื ืžืกืš 2025-11-29 ื‘-9.57.59](https://cdn-uploads.huggingface.co/production/uploads/69183cd79510e3441ef86afc/o4D5WklP1INIFBvvubdf3.png)
![ืฆื™ืœื•ื ืžืกืš 2025-11-29 ื‘-9.58.14](https://cdn-uploads.huggingface.co/production/uploads/69183cd79510e3441ef86afc/o9L76PgbO7hWmyEZIfQHL.png)
![ืฆื™ืœื•ื ืžืกืš 2025-11-29 ื‘-9.58.25](https://cdn-uploads.huggingface.co/production/uploads/69183cd79510e3441ef86afc/5BaZaCtq0RDU4Sg_kAneC.png)
![ืฆื™ืœื•ื ืžืกืš 2025-11-29 ื‘-9.58.36](https://cdn-uploads.huggingface.co/production/uploads/69183cd79510e3441ef86afc/0YggL_58zalfn50WokKf0.png)
![ืฆื™ืœื•ื ืžืกืš 2025-11-29 ื‘-9.58.48](https://cdn-uploads.huggingface.co/production/uploads/69183cd79510e3441ef86afc/S896FbQSOUX4pQTl5ym4d.png)
### 8.3 Model Evaluation
For each model we generated:
- `classification_report` (precision, recall, F1-score, support)
- Confusion matrix
- Interpretation of the types of errors the model makes
Below is a summary of the results:
#### **Logistic Regression**
- Achieved **perfect classification** on the test set (F1 = 1.00).
- The confusion matrix shows **0 errors**.
- This suggests the engineered features were highly separable.
#### **Random Forest Classifier**
- F1-score โ‰ˆ **0.79**
- Stronger recall for Class 0 (low delay), weaker for Class 1 (high delay).
- Confusion matrix shows the model tends to **miss high-delay flights** (false negatives).
#### **Gradient Boosting Classifier**
- F1-score โ‰ˆ **0.85**
- Better balance between precision and recall compared to Random Forest.
- Fewer false negatives than Random Forest and more consistent performance overall.
### 8.3 Which Model Performs Best โ€” and Why?
The **best model is the Logistic Regression**, because:
- It achieves **perfect predictive performance** on this dataset.
- It cleanly separates the engineered feature space into the two classes.
- It avoids the false negatives that are most critical in this task.
- Its confusion matrix shows **zero misclassifications**.
While this may indicate a highly separable dataset rather than model superiority alone,
within the scope of this assignment **it is the clear winner**.
---
### 8.4 Winner: Exporting and Uploading the Model
We exported the winning model (Logistic Regression) to a pickle file and uploaded it to the HuggingFace repository:
- **File:** `winning_classifier_model.pkl`
- Stored alongside the earlier regression winning model file:
- `winning_model.pkl`
Both files live in the same HuggingFace model repository as required.
# ๐ŸŽฅ 9. Video Presentation
Your recording should include:
* Quick dataset overview
* Key EDA takeaways
* How you encoded and engineered features
* Explanation of each model
* Confusion matrices
* Why Gradient Boosting won
* Summary of lessons learned