📊 Example Datasets

A collection of 61 polar questions in Persian, extracted from 1 hour of recordings of online gaming. Questions are annotated along five dimensions (Function, Intonation, Focus, Word order, Surprisal), presented with four levels of textual information (transcript, transliteration, translation, gloss), and enriched with three types of media content (video link, audio file, periogram image). Users are invited to familiarize themselves with the sound of real-life Persian intonation.

A collection of 1181 backchannel signals in Italian, extracted from a corpus of unscripted dyadic conversations. Backchannels are annotated along four dimensions (Function, Content, Distribution, Form) and presented with four levels of textual information (repetition, words, syllables, duration). Users are invited to use the Sidebar and Visualization options to focus on a subset of the collection and to modify the appearance of the plot.

A collection of vocalizations, calls and other sounds produced by 15 animals. Users are invited to listen to the sounds and drag the datapoints across the plot to categorize the sounds based on Danger (the average level of threat that the animal poses to humans) and Class (whether the animal is best described as bird, mammal or reptile). Users can code their Confidence level using colors.

!Spoiler alert! A possible solution for the AnimalVoices workbench, with names and images for each animal. Do not open before working on the AnimalVoices example!