Once you've created your project, defined your questions and uploaded the relevant data sets, the algorithms will fire away and produce an output of the statistically significant findings for you to explore. As you explore the output, you can select the most interesting, relevant and actionable discoveries to include on insight cards.
To view the results of the analysis, open a project and select 'View' next to the comparison you wish to explore.
You will initially be brought to the insights screen where any previously created insight cards will be displayed. To access the results of the comparison click 'Explore data'.
Use the filter options at the top of the screen to navigate through the different components of the analysis - we recommend starting with topics before moving into the more granular elements. You can toggle between different sides of the comparison by using the labelled tabs.
A list of linguistic features (topics, grammar, phrases, words, emotions) will display underneath the filters. In the example above, the list would include the differences in topics applicable to US gamers.
The measure displayed next to each feature will vary based on whether you are looking at differences, frequencies or similarities:
Differences – relative difference, a measure of how much more prevalent that particular feature is in one data set compared to the other
Frequencies & similarities – the percentage of each data set comprised of a particular feature, frequencies will contain all linguistic features present in the data sets while similarities will only show those with a relative difference between 0.9 and 1.1 (indicating a similar prevalence in the two data sets).
When you find something interesting and relevant to your research question, hover over the linguistic feature and click the lightbulb icon (to create a new insight) or the plus icon (to add to an existing insight).
Clicking into a linguistic feature from the list will reveal verbatim examples of it being used in the analyzed text. This can help to understand the context in which something is being used and determine whether it is insightful.