Auditing and Exploring Dielectric Measurements of Water

Imports, audits, compares and visualises dielectric measurements of water under fuel contamination and ageing. Provides descriptive summaries, frequency-response comparisons, confidence intervals based on physical-unit means, and training and testing splits that keep observations from the same experimental unit together. Experimental-unit safeguards address pseudoreplication concerns discussed by Hurlbert (1984) . Grouped splitting is consistent with the structured-data validation principles discussed by Roberts et al. (2017) . Student t confidence intervals use the standard mean-interval formula described in the NIST/SEMATECH e-Handbook of Statistical Methods < https://www.itl.nist.gov/div898/handbook/eda/section3/eda352.htm>. Experimental-unit identifiers are supplied by the user and are not inferred from material labels. This research prototype does not provide validated contaminant prediction or causal inference.

DielectricWaterR

An R package for auditing, exploring, and comparing dielectric measurements of fuel-contaminated water.

Overview

DielectricWaterR supports reproducible analysis of frequency-dependent dielectric measurements across fuel types, water media, concentrations, and ageing times.

Main features

• Import dielectric measurement data.
• Audit data quality and consistency.
• Visualise dielectric fingerprints.
• Compare dielectric fingerprints.
• Analyse frequency-response measurements.
• Audit observations excluded from frequency analysis.
• Compare matched frequencies across ageing times.

Installation

Install from a local package directory:

install.packages(
  '/path/to/DielectricWaterR',
  repos = NULL,
  type = 'source'
)

Example

library(DielectricWaterR)

# Supply your own authorised Excel workbook.
# The original experimental dataset is not distributed.
data_path <- "path/to/your_authorised_data.xlsx"

data <- read_dielectric(data_path)
audit <- audit_dielectric(data)

Scientific limitations

The package currently provides descriptive analysis. Matched-frequency comparisons do not establish causal effects or independent experimental replication. The package does not provide validated contaminant prediction or contaminant identification.

Version

0.2.0 — Initial research prototype.

Authors

Muhammad Zahir Khan (maintainer) and Luqman Abidoye

Flexible Excel import

Standard headers or explicit mappings can be used on any worksheet. Frequencies must already be in Hz and ages in hours.

x <- read_dielectric('my_data.xlsx', sheet = 'Measurements',
                     layout = 'standard')
x <- read_dielectric('my_data.xlsx', sheet = 'Measurements',
                     column_map = c(frequency_hz = 'Frequency',
                                    permittivity = 'Epsilon'))

Public tests use an entirely synthetic Excel workbook. The private experimental workbook is not distributed or required.