Elemental chemical compositions of float glass samples
datasetposted on 25.03.2020 by Soyoung Park, Alicia Carriquiry
Datasets usually provide raw data for analysis. This raw data often comes in spreadsheet form, but can be any collection of data, on which analysis can be performed.
We measured the elemental chemical composition of architectural float glass fragments using inductively coupled mass spectrometry with a laser ablation add-in (LA-ICP-MS). The unit for all measurements is parts per million (ppm).
Measurements of 18 elemental concentrations were obtained from each fragment at each measurement occasion. These data can be used for statistical analysis with the purpose of evaluating forensic trace evidence.
The data collection and measurement process in this database were carefully designed by the authors to enable understanding similarities and differences in elemental composition within a fragment, between fragments within a pane, between panes produced by the same manufacturer, and between manufacturers, to help in forensic glass evaluation. We received 48 panes that were produced on consecutive days, from two glass manufacturers in the U.S. Half of each pane was broken into small fragments and 24 fragments were randomly sampled from each half pane. To compute well-conditioned estimates of high-dimensional covariance matrices at all levels, we replicated measurements on each fragment; for three of the 24 fragments from a pane, we obtained 20 replicate measurements, and for the other 21 fragments, we made five replicate measurements.
Because of the design of the sample, the dataset will be a valuable resource for anyone working on methods for estimating and comparing covariance matrices, anyone who is interested in shrinkage methods and anyone who wishes to develop methods for the analyses of multivariate data.