Scribd
Upload
2 views5 pages

Program 3

The document outlines a machine learning pipeline for predicting delivery times using a dataset from Amazon. It includes steps for data loading, handling missing values, feature engineering, model training, evaluation, and saving the model. The final model achieved a mean absolute error of approximately 34.8 and an R2 score of 0.196.

Uploaded by

shilparn.fet.aiml
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
2 views5 pages

Program 3

The document outlines a machine learning pipeline for predicting delivery times using a dataset from Amazon. It includes steps for data loading, handling missing values, feature engineering, model training, evaluation, and saving the model. The final model achieved a mean absolute error of approximately 34.8 and an R2 score of 0.196.

Uploaded by

shilparn.fet.aiml
Copyright
© All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd

3/10/26, 10:49 PM Untitled32.

ipynb - Colab

# Step 1: Import Required Libraries

import pandas as pd
import numpy as np
import [Link] as plt
from sklearn.model_selection import train_test_split
from [Link] import LabelEncoder
from [Link] import StandardScaler
from [Link] import Pipeline
from [Link] import RandomForestRegressor
from [Link] import mean_absolute_error, r2_score
import joblib

# Step 2: Load Dataset

data = pd.read_csv("amazon_delivery.csv")
print("Dataset Loaded Successfully")
print([Link]())

Dataset Loaded Successfully


Order_ID Agent_Age Agent_Rating Store_Latitude Store_Longitude \
0 ialx566343618 37 4.9 22.745049 75.892471
1 akqg208421122 34 4.5 12.913041 77.683237
2 njpu434582536 23 4.4 12.914264 77.678400
3 rjto796129700 38 4.7 11.003669 76.976494
4 zguw716275638 32 4.6 12.972793 80.249982

Drop_Latitude Drop_Longitude Order_Date Order_Time Pickup_Time \


0 22.765049 75.912471 2022-03-19 11:30:00 11:45:00
1 13.043041 77.813237 2022-03-25 19:45:00 19:50:00
2 12.924264 77.688400 2022-03-19 08:30:00 08:45:00
3 11.053669 77.026494 2022-04-05 18:00:00 18:10:00
4 13.012793 80.289982 2022-03-26 13:30:00 13:45:00

Weather Traffic Vehicle Area Delivery_Time \


0 Sunny High motorcycle Urban 120
1 Stormy Jam scooter Metropolitian 165
2 Sandstorms Low motorcycle Urban 130
3 Sunny Medium motorcycle Metropolitian 105
4 Cloudy High scooter Metropolitian 150

Category
0 Clothing
1 Electronics
2 Sports
3 Cosmetics
4 Toys

[Link]

(43739, 16)

# Step 4: Check Missing Values

print("Missing Values in Dataset")


print([Link]().sum())

Missing Values in Dataset


Order_ID 0
Agent_Age 0
Agent_Rating 54
Store_Latitude 0
Store_Longitude 0
Drop_Latitude 0
Drop_Longitude 0
Order_Date 0
Order_Time 0
Pickup_Time 0
Weather 91
Traffic 0
Vehicle 0
[Link] 1/5
3/10/26, 10:49 PM [Link] - Colab
Area 0
Delivery_Time 0
Category 0
dtype: int64

# Step 5: Handle Missing Values

# Numeric columns
numeric_cols = data.select_dtypes(include=['int64','float64']).columns
data[numeric_cols] = data[numeric_cols].fillna(data[numeric_cols].mean())

# Categorical columns
categorical_cols = data.select_dtypes(include=['object']).columns

for col in categorical_cols:


data[col] = data[col].fillna(data[col].mode()[0])

print("Missing values handled")


print([Link]().sum())

Missing values handled


Order_ID 0
Agent_Age 0
Agent_Rating 0
Store_Latitude 0
Store_Longitude 0
Drop_Latitude 0
Drop_Longitude 0
Order_Date 0
Order_Time 0
Pickup_Time 0
Weather 0
Traffic 0
Vehicle 0
Area 0
Delivery_Time 0
Category 0
dtype: int64

# Step 6: Feature Engineering

# Create Age Group Feature


data['Age_Group'] = [Link](
data['Agent_Age'],
bins=[18,25,35,50],
labels=['Young','Adult','Senior']
)

# Convert Traffic category to numerical level


traffic_map = {
"Low":1,
"Medium":2,
"High":3,
"Jam":4
}

data['Traffic_Level'] = data['Traffic'].map(traffic_map)

print([Link]())

Order_ID Agent_Age Agent_Rating Store_Latitude Store_Longitude \


0 ialx566343618 37 4.9 22.745049 75.892471
1 akqg208421122 34 4.5 12.913041 77.683237
2 njpu434582536 23 4.4 12.914264 77.678400
3 rjto796129700 38 4.7 11.003669 76.976494
4 zguw716275638 32 4.6 12.972793 80.249982

Drop_Latitude Drop_Longitude Order_Date Order_Time Pickup_Time \


0 22.765049 75.912471 2022-03-19 11:30:00 11:45:00
1 13.043041 77.813237 2022-03-25 19:45:00 19:50:00
2 12.924264 77.688400 2022-03-19 08:30:00 08:45:00
3 11.053669 77.026494 2022-04-05 18:00:00 18:10:00
4 13.012793 80.289982 2022-03-26 13:30:00 13:45:00

[Link] 2/5
3/10/26, 10:49 PM [Link] - Colab

Weather Traffic Vehicle Area Delivery_Time \


0 Sunny High motorcycle Urban 120
1 Stormy Jam scooter Metropolitian 165
2 Sandstorms Low motorcycle Urban 130
3 Sunny Medium motorcycle Metropolitian 105
4 Cloudy High scooter Metropolitian 150

Category Age_Group Traffic_Level


0 Clothing Senior NaN
1 Electronics Adult NaN
2 Sports Young NaN
3 Cosmetics Senior NaN
4 Toys Adult NaN

# Step 7: Select Important Features

features = [
'Agent_Age',
'Agent_Rating',
'Traffic_Level'
]

X = data[features]
y = data['Delivery_Time']

# Step 8: Split Dataset

X_train, X_test, y_train, y_test = train_test_split(


X,
y,
test_size=0.2,
random_state=42
)

print("Training Data Size:", X_train.shape)


print("Testing Data Size:", X_test.shape)

Training Data Size: (34991, 3)


Testing Data Size: (8748, 3)

# Step 9: Create ML Pipeline

pipeline = Pipeline([
('scaler', StandardScaler()),
('model', RandomForestRegressor(n_estimators=100))
])

# Step 10: Train Model

[Link](X_train, y_train)

print("Model Training Completed")

/usr/local/lib/python3.12/dist-packages/sklearn/utils/[Link]: RuntimeWarning: invalid value enc


updated_mean = (last_sum + new_sum) / updated_sample_count
/usr/local/lib/python3.12/dist-packages/sklearn/utils/[Link]: RuntimeWarning: invalid value enc
T = new_sum / new_sample_count
/usr/local/lib/python3.12/dist-packages/sklearn/utils/[Link]: RuntimeWarning: invalid value enc
new_unnormalized_variance -= correction**2 / new_sample_count
Model Training Completed

# Step 11: Evaluate Model

predictions = [Link](X_test)

mae = mean_absolute_error(y_test, predictions)


r2 = r2_score(y_test, predictions)

print("Mean Absolute Error:", mae)


[Link] 3/5
3/10/26, 10:49 PM [Link] - Colab
print( Mean Absolute Error: , mae)
print("R2 Score:", r2)

Mean Absolute Error: 34.804068964512744


R2 Score: 0.19591032364467642

# Step 12: Visualization

[Link](y_test, predictions)
[Link]("Actual Delivery Time")
[Link]("Predicted Delivery Time")
[Link]("Actual vs Predicted Delivery Time")
[Link]()

# Step 13: Save Model

[Link](pipeline, "delivery_prediction_model.pkl")

print("Model saved successfully")

Model saved successfully

# Step 14: Load Model and Predict

model = [Link]("delivery_prediction_model.pkl")

sample_data = [[30,4.5,3]]

prediction = [Link](sample_data)

print("Predicted Delivery Time:", prediction)

Predicted Delivery Time: [127.56288703]


/usr/local/lib/python3.12/dist-packages/sklearn/utils/[Link]: UserWarning: X does not have v
[Link](

[Link] 4/5
3/10/26, 10:49 PM [Link] - Colab

[Link] 5/5

You might also like