|
|
|
dataCleaning
|
|
collection: CGMS
The tool aims at identifying duplicates with distinct determinations and this way assist in the identification of errors and/or new determinations. "Suspect" records are those that have the same name and collector number, but different information in the fields "genus + species + subspecies". This comparison is useful in detecting misspelling and duplicates with different determinations in the herbaria participating in the speciesLink network. Different collections have different standards for registering collectors’ names. To compare this field, the system recognizes variations in the order of the name and initials. For example, Barbosa, E. is recognized as the same as E. Barbosa. The system also identifies the primary collector when registered along with secondary collectors. For example Ferreira, C.A.C. is the same as Ferreira, C.A.C., Fiúza E. & Thomas, W.W. The list of possible duplicates is grouped by collector and collector number and can be ordered by family, genus and last update. The last updated is the date the record entered the speciesLink system. When a collection updates its data, new records present the date of the last update, but old ones present the date when they originally entered the system. This column can be used as a control as to which records already have been reviewed by the collection. Filters were created with the intention of identifying suspects records with more "critical" errors and this way contribute to the completeness and updating of taxonomical determinations, comparing them with duplicates of other herbaria. When clicking on the 'undetermined family' filter, the list shows the records that have the family determined in one collection and the absence of this information for in other collections. The same applies to 'undetermined genus'. Clicking on the item "same collection date" one restricts the list to records that have the same collector name and number and date of collection, which increases the probability that different herbaria records really are duplicates. |