End-user workflows for name matching

From TETTRIs
Jump to: navigation, search

The workflows that end-users follow vary significantly based on the specific use case, ranging from checking a single name to uploading a regional or monographic checklist with thousands of names. There are three main types of name-checking processes:

  1. Direct Use of the Aggregator's Name Matching Mechanisms: Utilizing the tools provided directly by the aggregator.
  2. Using Third-Party Tools: Leveraging tools such as [OpenRefine] that access the aggregator's name matching services.
  3. Using Local Tools: Downloading the aggregator's data and using local tools to perform the matching.

The choice of method mainly depends on the expected result but also on the number of records to be matched and on the technical in-house expertise available to the user. A type 3 process usually requires some expertise in biodiversity data management. TETTRIs provides links to download sites to get the aggregator's data. For type 2, TETTRIs will provide some example use cases that have been successfully tested. For type 1 (direct use of the aggregator’s services, the respective documentation will be pointed out in a list paralleling the list of general capabilities of aggregators.

The choice of method depends on the expected outcome, the volume of records to be matched, and the technical expertise available to the user. Type 3 processes generally require expertise in biodiversity data management. TETTRIs offers links to download sites for the aggregator's data. For type 2 processes, TETTRIs provides example use cases that have been successfully tested. For type 1 processes, relevant documentation will be documented paralleling the listed list of general capabilities of aggregators.

For the process itself, we can in principle distinguish 4 phases:

  • Preparing the data

In all cases, a list of names is needed in text-only format, which can be created from a spreadsheet column or be part of a table containing these names. One name only in one line is always required.

  • Submitting the data

Depends on the type of checking process.

  • Getting and interpreting the results

Essentially, for process type 1 and 2 the results are provided by listing exact matches and possible candidates, i.e. names that match the input to a certain extent. Interpretation refers to assessing the candidates and, if appropriate, selecting one of them as the correct match.

  • Incorporating the results locally

On the one hand, this refers to corrections of names made locally as a result of candidate matching. On the other hand, once matches were made unambiguously, this may result in incorporating the aggregator's name ID into the local dataset, to allow linkage to the aggregator and (if such functionality is made available) interaction with the aggregator.