TAXI DEMAND PREDICTION
Major Project - II (AR-6)
Batch 2021-25 | Sem - VIII
Status Report Part 1 | 25.02.2025
Mentor : Dr. Amit Kumar Rathi
Team Members -
Anubhav Chauhan - 211B057
Isha Sehgal- 211B141 1
INDEX
[Link]. Content Page no. [Link]. Content Page no.
9 Project Timeline 12
1 Overview 3
10 Outcome 13
2 Other Existing Models 4
11 References 14
3 Chosen Model 5
4 Block Diagram of the proposed 6
model
5 Use Case of the proposed model 7
6 Feasibility Study 8
7 Project Requirements 9-10
8 Techstack 11
2
Problem Statement
In urban environments, taxi service availability is often inconsistent,
leading to inefficiencies such as prolonged wait times for passengers and
revenue loss for drivers due to idle time or incorrect positioning. This
unpredictable demand often results in oversupply of taxis in low demand
areas or undersupply of taxis in high demand area.
Objective
OVERVIEW This project aims to develop a Machine Learning-based Taxi Demand
Prediction Model that forecasts the number of taxi rides required in
specific locations and timeframes by analyzing historical ride data, traffic
patterns, weather conditions, and other relevant factors.
Motivation
❖ The project will deliver a user-friendly prediction model optimize fleet
distribution, enhance urban mobility, and improve overall service
efficiency.
3
Some pre-existing models being employed in the market :
• Uber’s Demand Forecasting Model [1]
OTHER EXISTING • Lyft’s Predictive Demand System [2]
MODELS
• Grab’s AI-Based Demand Prediction [3]
• Ola’s Ride Demand Forecasting System [4]
4
• The ARIMA (AutoRegressive Integrated Moving Average) model is a
powerful statistical tool commonly used for time-series forecasting. It
is widely used for short term demand prediction.
CHOSEN MODEL: • In the context of taxi demand prediction, ARIMA can be applied to
ARIMA MODEL analyze past demand patterns to predict future taxi demand based on
trends, seasonality, and autocorrelations.
• ARIMA consists of three components- AutoRegression (AR) for past
values, Differencing (I) to make data stationary, and Moving Average
(MA) for past forecast errors.
5
Machine Learning Model
Figure 1: Block Diagram of the proposed model [5] 6
Taxi Demand Prediction
Taxi Demand
Figure 2: Use Case of the proposed model 7
FEASIBILITY STUDY
1. Data Availability: Availability of various datasets such as ‘Harvard Dataverse’,
‘Berkleys’ and ‘Kaggle’ for training and testing.
2. Resource Availability: Necessary hardware, software, and expertise are obtainable.
3. Operational Integration: Integration with existing infrastructure is achievable.
4. Technical Expertise: Required technical expertise is available or can be acquired .
8
PROJECT REQUIREMENTS
Requirements are the basic constraints that are required to develop a system.
Requirements are collected while designing the system. The following are the
requirements that are to be discussed:
Functional Non-Functional Environmental
Requirements Requirements Requirements
9
FUNCTIONAL NON - FUNCTIONAL ENVIRONMENTAL
REQUIREMENTS REQUIREMENTS REQUIREMENTS
.
• Data Collection and Processing • Scalability for Large Datasets Software Requirements:
• Programming Language
• Model Training and Prediction • Accuracy and Precision of (Python)
Predictions • Libraries and Frameworks
• Visualization of Prediction (NumPy, Pandas, Scikit-learn,
Results • Performance and Speed of TensorFlow)
Computations
• Data Storage and Retrieval Hardware Requirements:
• Data Security and Privacy • Processor (CPU/GPU)
• Validation of Model Accuracy • Memory (RAM)
• Storage (SSD/HDD)
10
TECHSTACK
11
PROJECT TIMELINE
DATA COLLECTION MODEL
& PREPROCESSING TRAINING
12
● The project aims to develop an efficient Machine
Learning model to accurately predict taxi demand
based on historical data, time, location, and
external factors.
OUTCOME
● This will help optimize driver allocation, improve
service availability, and reduce operational costs for
better ride-hailing efficiency.
13
REFERENCES
[1] [Link]
[2] [Link]
[3] [Link]
areas/
[4] [Link]
[5] [Link]
14
Thank you for your time.
15
Fig 3. Actual Temperature Prediction 16