Tools for Cleaning High-Frequency Real-Time Location Tracking Data.

trackclean

Tools for cleaning high-frequency real-time location tracking data.

trackclean was developed to process data from playground movement research, but applies to any study collecting high-frequency positional data from people moving within a defined space — classrooms, sports facilities, rehabilitation settings, and similar environments.

Installation

# Install from CRAN
install.packages("trackclean")

# Or install the development version from GitHub
# install.packages("devtools")
devtools::install_github("tomasbil/trackclean")

Example Data

The package includes a small example dataset that can be used to trial the full pipeline without any real data. It simulates 10 children tracked during a school recess on a 40m × 60m playground using a UWB positioning system.

library(trackclean)
library(readr)

raw_data   <- read_csv(system.file("extdata", "raw_tracking_data.csv", package = "trackclean"))
id_mapping <- system.file("extdata", "id_mapping.csv", package = "trackclean")

The example dataset includes:

• 10 participants with raw tag IDs 1–10, mapped to child IDs 5001–5010
• ~13.5 minutes of data (11:45:00–11:58:30), with observations both inside and outside the analysis window
• Sub-second timestamps causing multiple readings per second — handled by standardize_to_seconds()
• Randomly dropped seconds creating gaps — handled by interpolate_gaps()
• One tag replacement: participant 5003 starts on raw tag ID 3, which is swapped to raw tag ID 11 at 11:51:00 — handled by fix_tag_replacement()

Analysis parameters for this dataset:

Expected input format

Raw tracking data (raw_tracking_data.csv):

• ID: raw tag ID as assigned by the tracking system
• At: timestamp (POSIXct-readable, sub-second precision supported)
• X, Y: position in meters

ID mapping (id_mapping.csv):

• raw_id: tag ID as it appears in the raw data
• child_id: standardized participant ID to use in analysis
• A participant with a replaced tag appears twice (one row per tag, same child_id)

Quick Start

Optional: Fix Tag Replacements

If a participant's tag was replaced during data collection, run this before the main pipeline:

raw_data <- fix_tag_replacement(
  data = raw_data,
  original_id = 3,
  replacement_id = 11,
  replacement_time = "2025-03-18 11:51:00"
)

This will:

• Keep observations from tag 3 before 11:51
• Rename tag 11 observations from 11:51 onwards to tag 3
• Remove tag 3 observations from 11:51 onwards (duplicate/invalid)
• Remove tag 11 observations before 11:51 (not yet attached)

1. Prepare Your ID Mapping

Create a CSV file with two columns mapping raw device IDs to your participant IDs:

raw_id,child_id
1,5001
2,5002
3,5003

Or use the bundled example file:

id_mapping <- system.file("extdata", "id_mapping.csv", package = "trackclean")

2. Run the Complete Pipeline

library(trackclean)
library(readr)

raw_data <- read_csv(system.file("extdata", "raw_tracking_data.csv", package = "trackclean"))

# Fix tag replacement first (if applicable)
raw_data <- fix_tag_replacement(
  data = raw_data,
  original_id = 3,
  replacement_id = 11,
  replacement_time = "2025-03-18 11:51:00"
)

cleaned_data <- clean_playground_data(
  data = raw_data,
  id_mapping = system.file("extdata", "id_mapping.csv", package = "trackclean"),
  analyze_start = "2025-03-18 11:47:00",
  analyze_end   = "2025-03-18 11:57:00",
  bell_start    = "2025-03-18 11:53:00",
  bell_end      = "2025-03-18 11:58:00",
  output_file   = "cleaned_data.csv"
)

3. Use Individual Functions

For more control, run each step separately:

# Step 1: Map IDs
data <- map_ids(raw_data, id_mapping)

# Step 2: Mark time periods
data <- mark_time_periods(
  data,
  analyze_start = "2025-03-18 11:47:00",
  analyze_end   = "2025-03-18 11:57:00",
  bell_start    = "2025-03-18 11:53:00",
  bell_end      = "2025-03-18 11:58:00"
)

# Step 3: Standardize to seconds
data <- standardize_to_seconds(data)

# Step 4: Interpolate gaps
data <- interpolate_gaps(
  data,
  max_gap_small = 10,
  max_position_change = 0.3
)

Key Features

Two-Phase Gap Interpolation

The package uses a two-phase approach to handle missing data:

Phase 1: Interpolates small gaps (≤10 seconds by default)

• Uses linear interpolation between known points
• Appropriate for brief signal losses

Phase 2: Interpolates larger gaps conditionally

• Only when position change between endpoints is minimal (≤30cm by default)
• Indicates the participant remained stationary during the gap
• Prevents false movement estimates for longer signal dropouts

Quality Assurance

All functions provide:

• Progress messages and summaries
• Data integrity checks
• Row count validation
• Clear flagging of imputed vs. original data

Function Reference

Output Columns

The cleaned dataset includes these flags:

• id_code: Standardized participant ID
• Analyze: 1 if within analysis period, 0 otherwise
• Bell: 1 if within bell period, 0 otherwise (if specified)
• n_entries: Original number of signals in that second
• standardized: 1 if multiple signals were averaged, 0 otherwise
• imputed: 1 if row added via phase 1 interpolation
• imputed_large: 1 if row added via phase 2 interpolation

Parameters

Customizable Thresholds

cleaned_data <- clean_playground_data(
  data = raw_data,
  id_mapping = "id_mapping.csv",
  analyze_start = "2025-03-18 11:47:00",
  analyze_end   = "2025-03-18 11:57:00",
  max_gap_small = 5,             # Phase 1: ≤5 seconds
  max_gap_large = 30,            # Phase 2: ≤30 seconds max
  max_position_change = 0.5      # Phase 2: ≤50cm movement
)

Author

Tomas Bilevicius

License

CC BY 4.0 — you are free to use, share, and adapt this package for any purpose, including commercially, as long as you give appropriate credit to the author.