Qualichat Wiki

Welcome to the Qualichat Wiki! Here we summarize the Qualichat software features ongoing in Python. End-to-end, #OpenDialog #OpenScience for the community.

---

Interaction anchorage features (code: qc/anchor):

Interaction anchors are announcements for entries and exits in a given thematic frame. The technical features of mediated groups can delimit the textual structure of a linguistic corpus to be analyzed by natural language processing.

Authoritarian index feature (code: qc/index):

Public opinion in a group can be measured by authoritarian, pluralist, and transcendental typological levels. With this code, the investigator and user will identify the levels of authoritarianism summed up in the amounts of text, links and symbols used, and the speed of interaction.

Human and non-human sharing patterns identification features (code: qc/bots):

Non-human presence in mediatized groups is more common than we imagine. Qualichat aims to democratize free group discussions by allowing users and researchers to take a true X-ray of the presence of bots in groups. Human patterns are contrasted with bot patterns at a single code.

Shared news features(code: qc/news):

Instant messaging companies have been changing their link-sharing policies, but much damage has already been done. News shared in groups can be identified with this Qualichat feature. Ideological political polarizations may be perceived by researchers and group members.

Text structure features (nominal, verbal, and adjectival syntagms) (code: qc/nouns, qc/verbs, qc/adjectives):

Linguistic syntagmas refer to a group of related linguistic elements in a sentence. It could be by nouns, verbs, and adjectives, meaning a syntactic unit composed of one or more words that form sentences. With this feature, word graphs are generated from a dataset processed in Python. Also, in this package, the searcher will understand the syntagms connected to it by choosing a keyword. To get this, use qc/nouns/keyword, qc/verbs/keyword, and qc/adjectives/keyword.

User Group Constellation Features (code: qc/actors):

Actor constellation is a way to understand how users are intertwined through a debate. Different levels of engagement can be observed by the depth of conversation between them. Not necessarily an actor who talks the most is the most influential in the group. Still, a quantitative view on this can help in correlations with ethnographic immersion experiences. With this feature, phone numbers associated with a message line can be anonymized with pre-set lists, such as city names, songs, Brazilian names, pictures, plants. Note, Qualichat initially works only with natural language processing in Brazilian Portuguese, so local cultural aspects are clearly embedded in this code. When choosing a code qc/actors/name/cities from a list of cities, Brazilian cities start to name the actors instead of the phone number.

Thematic keyword searching features (code: qc/keyword/x/)

Frames of relevance are tastes for a particular debated topic. Keywords help the researcher and the user interested in the most discussed themes. From this code, graphics are generated with the feature variables found in this Wiki. Example. If you want to see theme variations per topic insert qc/keyword/time/x/y.

Interaction Timing features (code: qc/time)

Time defines the speed of interaction. Quantities of messages with super-fast, fast, regular, long, short, and long intervals can be made with this code. This code is one variable that can be associated with other ones, such as the constellation of actors, news, anchorages, keywords, i.a.. Example: qc/time/actors/actor1.

Quantitative data visualization features by frame analysis (code: qc/charts/chartype/v1, v2

Qualichat understands those different ways of visualizing quantitative data can help get qualitative insights. By joining two or multiple code variables, the following graphics can be applied: Line Graphs, Multi-Line Graphs Bar Graphs, Pie Charts, Mosaic Charts, Pyramids, Spider Charts, Waterfall Charts, Scatter Plots, Trellis Plots, Sunburst Charts, Stacked Area Charts, Stacked Bar Graphs, Trellis Bar Graphs, Multi-level Pie Charts, Scattergrams, Trellis Line Graphs, Radar Charts, Stacked Bar Graphs, and Scatter-Line Combo. Also, History Timelines e Frames Timeline are uniquenesses that can make the qualitative insights richer.

Sentiments features (code: qc/sentiments/)

Laughter, stickers, and emojis can be detected with a simple code command. From a previous list of emotion scales, researchers and users will identify which sentiments are most present in the groups and related variables, either by actor or topic. Searching for emotion symbols related to code variables can also be done.

Data Preprocessing features (code: qc/data/)

This is one of Qualichat's most significant innovations. Data organization from a txt file can be done automatically, correcting line break errors and eventual discontinuities.