Improved Data Analyzing on the NLG Platform
Classifying data values and gaining overview over specific data fields are basic requirements for producing high-quality automated text. Hence we have restructred the processes for analyzing data on the NLG platform:
Now you can adapt the analyses more precisely to the projects and their data (sub)sets and you also get more details about the statistics of your collections. In addition the creation of nodes in the Analyze tab is much easier now.
New analysis features
The analysis (previously also called histogram) of the data is carried out in the Analyze tab of the COMPOSER as before. The new features are:
- The analyses are now assigned to a project, no longer just to a collection.
- You have the possibility to set general filters (language and collection) to confine the analysis before starting it.
- You can now use multiple analyses for one project and access them via the name you have assigned to them.
- More detailed statistical information for each data set is provided by default.
- Searching for documents with the same value is possible now.
- You access the results of the analyses in the Transform tab by creating analysis result nodes
Create a data node with one click
If you want to use a data field as data node you just have to click on the button create node. The software recognizes the data type (e.g. string or number) and creates a data node with the field name as name. For further configuration you can access this new data node in the transform tab
When can I start using this feature? What does this mean to existing projects?
The new features are live monday 6th of July, it will automatically replace the existing “analzye” view, so you don’t need to change anything.
The existing histograms that have been created until now have been migrated automatically, and histogram_data() calls continue to work.