Uber Taxi Fares Project

Project Overview

The Uber Taxi Fares project aims to analyze, explore, clean, and structure the data, and then fit a linear model to understand the relationship between features for better business insights.

Dataset Overview

The dataset contains 200,000 samples, each with 7 characteristics. All of them have a data type of float except the pickup_datetime and id, which are of type int.

Preparing Data for EDA

We dropped 1 duplicate column, removed a sample because of missing values, and created the distance_km column as it was not provided.

Exploratory Data Analysis Findings

We created different visualizations to understand data distribution and characteristics. Some of the findings are:

• The majority of trips have a distance of less than 5 km.
• The fare amount for most trips is between 0 to 20 dollars.
• In the majority of trips, the number of passengers was between 1-6.
• Analysis of trips and revenue per month shows that summer months have fewer trips compared to the whole year.
• The number of trips per day is high on Monday and low on Saturday and Sunday.
• Analyzing the revenue per day shows that Monday has the lowest revenue compared to Friday.
• The mean fare amount with respect to the number of passengers shows no relationship, so we conducted a hypothesis test, which revealed a relationship.

Linear Regression Model

To fit the linear regression model, we first removed outliers and engineered some features.

Outliers

We removed outliers from the fare_amount, distance_km, and passenger_count columns.

Feature Engineering

• Created day and month columns from the pickup_datetime column.
• Created a rush_hour column indicating 1 for trips between 6 am-10 am and 4 pm-8 pm.
• Scaled all the variables.

Fitting the LR Model

Before fitting the model, we dropped latitude, longitude, pickup_datetime, day, and month and used only passenger_count, distance_km, and rush_hour.

• After finding the optimal hyperparameters using the grid search method, the model achieved an R² score of 0.61, indicating it explains 61% of the variance in the data.
• Analyzing the coefficients, we concluded:
  • For every 1.42 km traveled, the fare increases by a mean of 2.95 dollars or for every 1 km traveled, the fare increases by a mean of 2.08 dollars.

Predictive Model

After fitting the LR model, we fitted other models to find a predictor for real-time prediction. We engineered additional columns like daytime, nighttime, am_rush, pm_rush, etc.

• Applied Random Forest and XGBoost to the data and found that XGBoost outperforms the others with a variance of 0.62, which will be used for prediction purposes.

Libraries Used

• numpy
• pandas
• matplotlib
• seaborn
• scipy
• sklearn
• xgboost

Note

This is the end of the project. If you liked it or found it useful, please upvote it. If you have any suggestions, please let me know.