Multiview Visualisation of Uber Ride Booking Behaviour Data

This repository contains a project for Information Visualization.
The project explores Uber ride booking behaviour using three interactive multiview systems built with Python and Altair.

Project Overview

We designed and evaluated three visualisation systems to support exploratory analysis of ride-hailing data.
The systems focus on:

• ride demand across vehicle types
• temporal demand patterns
• relationship between ride distance and booking value
• booking outcomes
• interactive subset exploration

One of the systems also implements generalised selection through a temporal hierarchy.

Systems

System A — Demand Overview Dashboard

A standard multiview dashboard for fast and intuitive exploration.

Main features:

• bookings by vehicle type
• hourly demand patterns
• booking outcomes
• distance vs. booking value scatter plot
• click-based linked filtering

System B — Density and Time Characterisation System

A more analytical system using density-based visualisation and hierarchical interaction.

Main features:

• distance vs. booking value heatmap
• hourly booking status distribution
• generalised selection by time of day
• brushing and coordinated filtering

System C — Spatial and Temporal Ride Demand Dashboard

A location-focused system for origin–destination and temporal exploration.

Main features:

• top pickup locations
• top drop locations
• hour vs. weekday heatmap
• vehicle type filtering
• linked subset exploration

Dataset

Dataset source:
Uber Ride Analytics Dashboard Dataset on Kaggle

The dataset contains 150,000 ride booking records and 21 attributes, including:

• temporal attributes: date, time, hour, weekday
• categorical attributes: vehicle type, booking status, pickup and drop location
• quantitative attributes: ride distance, booking value, ratings

Tasks Supported

The systems were designed to support the following tasks:

• T1: Compare ride demand across vehicle types
• T2: Analyse temporal patterns in ride demand
• T3: Explore the relationship between ride distance and booking value
• T4: Investigate booking outcomes and ride completion patterns
• T5: Select and explore subsets of rides

Generalised Selection

System B implements generalised selection using a temporal hierarchy:

• Hour of day
• Time-of-day category:
  • Night
  • Morning
  • Afternoon
  • Evening

This allows users to move from detailed hourly analysis to broader temporal groupings.

User Evaluation

We conducted a user evaluation with 5 participants and compared the three systems using:

• task completion time
• number of errors
• user preference

Summary of findings:

• System A was the fastest and easiest to use
• System B provided stronger analytical depth
• System C offered the strongest spatial insight, but was more complex to use

Files

• report/ — final report PDF
• SystemA/ — source code for System A
• SystemB/ — source code for System B
• SystemC/ — source code for System C
• images/ — screenshots used in this README

Technologies Used

• Python
• Altair
• Pandas
• Jupyter Notebook