Data Evaluation, Transformation and Reporting

This distillation process is usually carried out using statistical procedures. However, as Sir Ronald Fisher remarked, for this process to be successful, the data set has to contain the information sought. 

No amount of elegant statistics or chemometric procedures will rescue inadequate or unreliable data. 

From an analytical viewpoint, statistical approaches can be subdivided into two types Exploratory Data Analysis (EDA) and Confirmatory Data Analysis (CDA). Exploratory data analysis is concerned with pictorial methods for visualising data shape and for looking for patterns in multivariate data. It should always be used as a precursor for selection of appropriate statistical tools to confirm or quantify, which is the province of confirmatory data analysis. CDA is about applying specific tools to a problem, quantifying underlying effects and data modelling. This is the more familiar area of statistics to the analytical community. 

Some basic aspects of EDA will be explored in Section 8.1, and in Section 8.2 the most frequently used CDA technique, linear regression analysis for calibration, will be covered. It is not intended to provide a statistical recipe approach to be slavishly followed. The examples used and references quoted are intended to guide rather than to prescribe. 

---

Data Evaluation, Transformation and Reporting Table 17 Calculated residuals... 

White et al. [53] reported open-access LC-MS cf. Ch. 9.2) for the rapid characterization of proteins. The system allows web-based sample submission and registration, automated data processing, interpretation, and report generation. The submitted amino-acid sequence is used to set parameters for the transformation of the mass spectmm into the protein molecular weight and to evaluate whether the submitted protein is actually detected. Data acquisition is performed on a TOF instrument in order to achieve sufficient mass accuracy. 

Lastly, modern analytical instruments are almost always interfaced with personal computers to provide sophisticated system control and the storage, treatment (for example the performance of Fourier transforms or calculations of derivative spectra) and reporting of data. Such systems can also evaluate the results statistically, and compare the analytical results with data libraries in order to match spectral and other information. All these facilities are now available from low-cost computers operating at high speeds. Also important is the use of intelligent instruments, which incorporate automatic set-up and fault diagnosis and can perform optimization processes (see Chapter 7). 

Even more recently, Burmester reported on the use of pyrolysis mass spectroscopy in lacquer studies (47,48). The results, when used with multivariate data analysis, prove to be a helpful provenance tool. Burmester has also extended the IR work through the use of a Fourier transform instrument and, further, evaluated the efficacy of using carbon-13 NMR measurements (49). 

---