README.md

metadata

language: en
license: mit
tags:
  - mobility
  - telecommunications
  - synthetic-data
  - 5g
  - network-optimization
datasets:
  - mobility-pattern

Mobility Pattern Dataset

Dataset Description

• Repository: mobility-prediction/dataset
• Paper: N/A
• Point of Contact: hello@unifyair.com

Dataset Summary

This dataset contains synthetic mobility patterns and network performance metrics for 100 users over a 3-day period. The data simulates realistic user movement patterns in a cellular network environment, including various mobility types, signal strengths, and network conditions.

Supported Tasks and Leaderboards

• Task 1: Mobility Pattern Prediction
• Task 2: Network Performance Optimization
• Task 3: Handover Decision Making

Languages

English

Dataset Structure

Data Instances

Each data instance represents a single measurement point for a user, containing:

• Timestamp
• User ID
• Spatial coordinates (x, y)
• Velocity and heading
• Connected cell information
• Signal strength
• Handover information
• Pattern type
• Network conditions

Data Fields

• timestamp: DateTime - Time of measurement
• user_id: String - Unique identifier for each user
• x: Float - X-coordinate in meters
• y: Float - Y-coordinate in meters
• velocity: Float - Movement speed in m/s
• heading: Float - Direction of movement in radians
• connected_cell: String - ID of the currently connected cell tower
• signal_strength: Float - Signal strength in dBm
• handover_needed: Boolean - Whether a handover is needed
• handover_target: String - Target cell for handover if needed
• pattern_type: String - Type of mobility pattern ('commuter', 'random_walk', 'stationary', 'high_mobility')
• network_load: Float - Network congestion level (0-1)
• sinr: Float - Signal to Interference plus Noise Ratio
• throughput_mbps: Float - Network throughput in Mbps
• device_type: String - UE capability category
• handover_latency: Float - Handover latency in milliseconds
• handover_success: Boolean - Whether the handover was successful

Data Splits

The dataset is provided as a single split containing 3 days of continuous data.

Dataset Creation

Curation Rationale

This synthetic dataset was created to facilitate research in mobility prediction and network optimization. It simulates realistic user movement patterns and network conditions that would be encountered in a real cellular network environment.

Considerations for Using the Data

Social Impact of Dataset

This dataset can be used to:

• Develop and test mobility prediction algorithms
• Optimize network resource allocation
• Improve handover decision making
• Train machine learning models for network optimization