Services > Data mining

 

Whatever your problem, there is often a significant amount of information already available in the  literature which can be critically reviewed to give you a head-start in a new programme. Bedrock Geosciences are experienced in the art of data mining. Although the process of ‘data mining’ in this manner appears to be relatively straightforward, this is not the case as is evidenced from the numerous so-called reviews which are no more than a collection of everything written on the topic – whether they are of relevance or not. Data mining in this manner needs to be carried out by focussed and experienced staff who understand the boundary conditions under which they are working and who are capable of a true, critical evaluation of the data. Effective data mining requires

1. Fully documented reports which include as much of the raw data as possible, as well as the original interpretation. Where possible, the raw data should include uncertainty values
2. If not available in the reports, a fully quality assured raw data set should be accessible
3. Clear indication of the status of any models used (eg have they been tested previously)
4. It is an advantage if the original project team can still be contacted and questioned
5. A sample depository is also of use, allowing re-analysis of material if necessary
6. Where possible, any new information should be compared with relevant laboratory, in situ (URL) and natural analogue data to produce an integrated whole, something which has rarely been done within any existing SAs
7. Where possible, any re-analysis should be integrated into an ongoing safety case or, at least, the new information should be placed in the context of an existing safety assessment (SA)

With such information at your finger tips you can then decide whether to work directly from the existing literature, rework the original data in the existing literature or initiate a new, focussed study.