Uber-Data-Analysis

This project focuses on analyzing Uber ride data to uncover insights and build predictive models for pricing. We leverage exploratory data analysis (EDA), feature engineering, and multiple regression algorithms to determine the best-performing model for predicting Uber ride prices.

---

📌 Problem Statement

We analyze Uber ride data available in tabular format, applying machine learning algorithms and statistical techniques to:

• Understand the relationships between features such as date, time, and location.
• Identify the most important variables affecting ride prices.
• Help Uber enhance its services by highlighting influential factors in pricing strategy.

---

🧰 Libraries Used

Left Side Libraries	Right Side Libraries
numpy	xgboost
pandas	lightgbm
matplotlib.pyplot	catboost
seaborn	warnings
sklearn.model_selection	sklearn.linear_model
sklearn.preprocessing	sklearn.ensemble
sklearn.metrics

---

📊 Project Workflow

Step	Description
Defining the Problem	Define the problem and project objectives.
Loading the Dataset	Import and load the Uber ride data.
Basic Cleaning	Clean the dataset by handling missing values.
Exploratory Data Analysis (EDA)	Analyze and visualize the data.
Univariate and Bivariate Analysis	Investigate individual and pairwise relationships between variables.
Correlation Analysis	Analyze correlations between features.
Feature Engineering	Create new features or modify existing ones.
Preprocessing	Prepare data for modeling.
Encoding Categorical Features	Convert categorical data into numerical format.
Scaling Numerical Features	Normalize or standardize numerical features.
Model Building	Build various regression models.
Linear Regression	Apply linear regression.
Random Forest	Implement random forest model.
XGBoost	Implement XGBoost model.
LightGBM	Implement LightGBM model.
CatBoost	Implement CatBoost model.
Model Evaluation	Evaluate models using RMSE and R² score.
Conclusion	CatBoost Regressor emerged as the best-performing model.

---

📊 Load Dataset

Sample: The dataset contains 300 rows and 46 columns. Below is a statistical summary of the columns:

Column	Count	Mean	Std	Min	25%	50%	75%	Max
timestamp	300	1.541233e+09	3.293688e+06	1.540000e+09	1.540000e+09	1.540000e+09	1.540000e+09	1.550000e+09
hour	300	11.31	7.27	0	4	12	18	23
day	300	17.72	10.09	1	13	17	28	30
month	300	11.58	0.49	11	11	12	12	12
price	276	15.97	8.42	3	9.5	13.5	22.5	38.5
distance	300	2.20	0.97	0.44	1.17	2.44	3.05	4.43
surge_multiplier	300	1.00	0.03	1	1	1	1	1.25
latitude	300	42.33	0.05	42.21	42.34	42.35	42.36	42.37
longitude	300	-71.07	0.02	-71.11	-71.08	-71.06	-71.05	-71.03
temperature	300	39.80	7.20	18.91	37.26	40.87	43.91	57.22
...	...	...	...	...	...	...	...	...

Note: The dataset contains various features related to Uber ride data, including price, distance, surge multiplier, temperature, and more.

---

Exploratory Data Analysis (EDA)

✅ Correlation matrix:

---

Model Evaluation

✅ Predict and Evaluate

Below are the performance metrics (RMSE and R² Score) for each regression model used:

Model	RMSE	R² Score
Linear Regression	7.289921	0.264664
Random Forest	2.833321	0.888921
XGBoost	2.445294	0.917263
LightGBM	2.527042	0.911638
CatBoost	3.007453	0.874848

📌 XGBoost achieved the lowest RMSE and highest R² score, making it the best-performing model in this experiment.

✅ Visualize comparison:

---

📈 Conclusion

From our model comparisons, CatBoost Regressor delivered the best performance based on evaluation metrics like RMSE and R² Score.
It is recommended for future Uber price prediction tasks due to its robustness and superior accuracy.

---

📄 License

This project is protected under copyright © Md Altamash Alam, 2025.

All rights reserved. Unauthorized copying, distribution, modification, or use of any part of this project without explicit permission is strictly prohibited.

If you wish to use or reference any part of this project for academic, personal, or commercial purposes, please contact the author for permission.

---

© Md Altamash Alam, 2025 – All Rights Reserved.