Exploratory analysis explores possible patterns or trends in the chemical data. Such methods can be done in a supervised/unsupervised manner and they can reduce large complex data sets into a series of optimized, simple and interpretable output. Below are list of activities in which we have used data exploratory methods:
- To assign a suspect sample to the predefined categories (which could be geographical origin, cultivar variety, conventional and organic production type, harvesting year or environmental condition, food fraud/adulteration and evaluation of health claim)
- To explore changes that are happening to the samples in the treatment chains or process control industry (such as differential analysis between influent and effluent wastewater in wastewater treatment plants)
- To discover generation of new transformation products of a parent compound via trend analysis under a treatment method (such as ozonation, photodegradation etc.)
- To study the daily/seasonal trend of emerging contaminants occurrence in river samples
- To extract significant features from mass spectral data like markers
For these purposes, we usually use clustering and classification methods that are used to measure the similarity or dissimilarity among samples. We have developed several software to implement such activities in our group.