Closed kodymoodley closed 7 months ago
I actually found that this is not actually a bug, it is just that the input file must be processed in Orange before connecting to the Actor Analysis widget. I.e., the columns of the input CSV get parsed and assigned a data type automatically e.g. "categorical". Actor analysis widget requires ALL CSV columns are of type "text". The solution is to use the "Edit Domain" widget to change the types of all columns to text. @f-hafner
Here is a test CSV file to use: dict.csv
For story data, you have to find Dutch text on the net (I cannot share the stories I have since there is personal data involved).
Tested and works in version 0.0.11
Story Navigator version
0.0.10
Orange version
3.36.2
Expected behavior
Bar plot displays grouped counts of the different halliday action counts
Actual behavior
Nothing is displayed in the bar plot
Steps to reproduce the behavior
Import set of textual stories using the Import documents widget and connect it to the Corpus input of the Actor analysis widget.
Import (using import CSV widget) a CSV file with first column as a list of words, and all subsequent columns are different groups of word categorisation labels for the words in column 1. Connect the output of that widget to the token categories input of Actor analysis widget.
Connect the custom token frequencies output of the Actor analysis widget to the data input of a bar plot widget
Additional info (worksheets, data, screenshots, ...)