Displayr

Please bring back the buttons on the text analysis!!! Specifically, the "categorize" button at the top. It is now much more difficult to categorize one response as multiple themes. To have to manually type in each category is slowing down efficiency by 10000000. I also now can't just select more than one response and select a theme and press "categorize" -- I have to go one by one. Please bring back the categorize button!!!

In the text classification module, we remove "chips" or themes from a response by clicking into the response and then hitting the x on the chip. The x's should be present without going into the detailed view for the response - that is, in the column next to the response - so that they can be edited quickly. If multiple responses are selected, then the relevant theme should be removed from all selected responses. If a theme that's being removed is not associated with a response, and that response is selected, then do not edit that response, but allow the other selected responses to nevertheless have the theme removed.

Ideally, when “navigating” in manual classification mode like this, it should lock the current response to a position where the drop-down always appears below the response, rather than variably appearing above or below it depending on screen real estate. It gets confusing when you don’t know whether you’re reading the response above or the response below, and have to re-adjust depending on how far up or down the visible list of responses you are.

When classifying text manually, we currently have to press enter to exit the drop-down, and then enter again to open the drop-down for the next response. If selecting a single theme for classification, this results in 3 taps on Enter for each response. However, as the current feature moves to the next response down and selects it (good), it should also automatically open the classification drop-down. To stop classifying, hit Esc.

When using the drag and drop feature in the Text Classification module, we used to be able to select responses, select one or more themes, and then click MOVE. Currently, we can drag and drop responses directly into a theme which is an improvement over-all, but this is solely an additive action. That is, we cannot MOVE responses into themes. Instead, we currently have to: Classify the response(s) to its desired theme. Right-click the responses. Remove the undesired themes from the drop-down. An improvement here would be to MOVE responses by dragging and dropping. When dragging, the user holds down SHIFT of CTRL on their keyboard and a tooltip appears on the mouse pointer to say "Move" When the mouse button is released on a theme, the responses get automatically removed from any other themes they've been assigned to, and classified only with the target theme.

In the Text Categorization interface, I often browse through comments looking for quotes to feature. While I can copy quotes by selecting and using copy-C on my keyboard, it took me a while to realize this was possible. It would be beneficial to have a more intuitive option, such as adding a 'copy' option to the right-click menu. This would enhance usability and make the feature more discoverable.

We are looking for a way to edit and visualize text categorizations in a format similar to the Edit Categorization tool, rather than the current table format. This would help us create more digestible tables for our community to understand better. It would be beneficial if Displayr could provide a function to edit and display categorizations in this manner.

Not doing so creates an extra ckick/task for the user in most cases, I would presume.

Currently the Create QCodes File automation only works with Binary - Multi variable sets or individual Nominal variables. However, this means that a common way text responses are collected is unable to use this. If you have a survey with multiple mentions, such as brands listed, stored across multiple text variables, it's most common to work with it as a Nominal - Multi or Binary - Multi (Compact) variable set, where each variable is associated with separate text variables and the categorized categories are the categories of each variable. Currently if you export each nominal variable individually as a QCodes file, you'll get multiple QCodes files that may not align. Instead, this automation needs to basically create a single, shared QCodes file that works across all the Nominal variables in the variable set to assemble one, shared code frame which can then be exported.

Incorporate the coding list of items into Text Categorization if a user selects Only One Theme, for example, brand lists, etc. The customer has said what they would like is when using the text categorization, the AI coding just gave them themes, rather than coding the actual people. I then tried the 'list of items' but this didn't seem to create a codeframe as such, and was not easily editable in terms of combining multiple categories. They would like the same as a' list of items' but for it to work the same as the AI/theme coding - but instead of themes, to code the individual items - e.g. people, brands, etc.