Wiki

Clone wiki

main / use_cases

Use Case Sketches

Transcribing an Interview

Transcribing a new interview

  1. In the project explorer view, Abbie clicks on the "New interview" button.
  2. A New Interview wizard appears. Abbie selects the participant(s) involved with the interview, enters the date of the interview, and provides a short (unique) description and an optional long description.
  3. Abbie also indicates in the wizard the path of the audio file containing the interview. Only wav. and mp3 files are currently supported.
  4. Abbie clicks on finish.
  5. A dialog appears with four tips:
    1. Press Ctrl-W to play/pause the audio file
    2. Press Ctrl-Q to rewind 5 seconds (See Preferences to change the time).
    3. Press Ctrl-E to fast forward 5 seconds.
    4. Press Ctrl-T to create a timestamp at the currently selected paragraph.
  6. Then, an interview is created under the directory Interviews in the project explorer. The name of the interview followed by the ids of the participants involved in the interview are displayed.
  7. A Text Editor appears. On top of the editor, there is a series of button to interact with the audio file (play, pause, rewind, fast forward, stop).
  8. Abbie starts the audio file and transcribe the interview. Each time she creates a new paragraph, the current time on the audio file is automatically associated with the paragraph.
    1. A green marker on the right of the editor shows that there is a timestamp.
    2. By hovering the mouse over the marker, the time is displayed.
    3. When clicking on the marker, the audio file plays at the selected timestamp.
  9. The editor is automatically saved every minute (configurable).
  10. Abbie wants to take a break, so she clicks on pause and closes Qualyzer.
  11. Later on, she reopens Qualyzer and the interview. The audio file cursor is set at the same timestamp where the pause occurred.

Producing an Interview Summary

TBD

Analyzing an Interview

Writing a Memo

TBD

Coding an interview for the first time

The first codes

  1. Abbie double-clicks on an interview. The interview is displayed in a text editor.
  2. Abbie reads the interview. She finds a particular sequence of words interesting and highlight it with her cursor then presses Ctrl-K (or Add Code in the context menu).
  3. A small dialog appears with two text boxes: one single-line text box labeled Code and the other, a multi-line text box labeled Definition.
  4. Abbie enters the words "Junk Food"
  5. A warning sign appears in the dialog: "This code does not exist and will be created when you click OK"
  6. Abbie enters a definition in the definition text box and presses ok.
  7. A marker appears on the left and the coded text is now lightly highlighted.
  8. When Abbie hover her mouse over the text, she sees a tooltip "Junk Food"
  9. Abbie continue to analyze the interview and enter new codes.

Reusing codes

  1. At some point, Abbie reads another sentence related to Junk Food. She highlight the text and press Ctrl-K.
  2. In the dialog, she enters "ju" and a small listbox appears with a list of codes: Junk Food, Just Right.
  3. She selects Junk Food and the definition of the code appears in the definition text box.
  4. Abbie clicks on ok.
  5. The text is now assigned to the Junk Food code.

Assigning multiple codes

  1. Abbie rereads the interview and think that a sentence does not only talk about junk food, but also about software bugs.
  2. Abbie select the part of the sentence that is related to software bugs and press Ctrl-K.
  3. A dialog appears. She select Software Bugs with the auto-complete feature and clicks on OK.
  4. When she hovers her mouse over the text on or the marker, she sees one tooltip with two codes in it: Junk Food, Software Bugs.

Modifying codes

  1. Abbie rereads the interview and think that a sentence does not really represent Junk Food. She right click somewhere in the sentence and select "Delete Code" in the context menu.
  2. The code is removed from the sentence, the marker is removed and the sentence is no longer highlighted. do we support undo? --bart
  3. Then, Abbie believes that the definition of Junk Food is not quite right. She go to the Code view, select the Junk Food code and double click on it (she could also have selected "Edit Code" in the context menu).
  4. A dialog appears with the code name and the definition. Abbie changes the definition and clicks on OK.
  5. The definition is now changed. Abbie could also have changed the name of the code, e.g., to fix a typo.

Alternative to Modifying Codes

  1. Abbie believes that the definition of Junk Food is not quite right. She finds a sentence coded with Junk Food, right clicks somewhere in the sentence, and selects the View Code option.
  2. A dialog appears with the codes associated to this sentence. Abbie selects the Junk Food code and the dialog is refreshed with the definition.
  3. Abbie changes the definition and clicks on OK.
  4. The definition is now changed.

Editing a coded interview

  1. While analyzing an interview Abbie realizes that some part of the interview was not transcribed.
  2. She clicks on the time marker (on the right) nearest to the place where she thinks the transcription is incomplete. The audio file plays.
  3. She starts to transcribe the missing parts. At some point, she saves the interview.
  4. The annotations (codes) that are placed after the edit are updated.
  5. Then, Abbie fixes some typos in sentences: some are highlighted (coded) some are not.
  6. She saves the interview and the coded sections are updated to take into account the edits.

This part is tricky. I don't know how easy it is to reconcile the model while editing or during each save. As long as we support basic editing of coded interviews, we should go with the easiest solution to implement. In the worst case, we could take a snapshot before the edit and if the user is not happy with our automatic update, he can revert the edits easily. --bart

Organizing codes

  1. Abbie wants to group some of the codes she created in abstract category.
  2. She opens the code hierarchy view. On the left, there is a flat list of all the codes. Beside each code, there is a number in parenthesis (i.e., the number of text sections associated with this code) On the right, there is a (empty) tree view. When she hovers her mouse over a code, a tooltip appears with the definition of the code.
  3. Abbie sees that she can group two codes, Junk Food, McDonald Food, under the common category Fast Food.
  4. Abbie creates a new code called Fast Food in the list view (she right-click and select new code in the context menu. She enters the code name and a definition).
  5. Then, Abbie drags and drops the Fast Food code from the list view to the tree view.
  6. Then, Abbie drags the Junk Food and McDonald Food codes from the list view and drops them on the Fast Food code in the tree view. The Junk Food and McDonald Food codes are now displayed under the Fast Food code.
  7. In the list view, the labels read like this: Junk Food (10), McDonald Food (5), Fast Food (0,15). The last label means that there are no text sections explicitly coded with Fast Food, but fifteen sections are coded with children of Fast Food.

The behavior is as follow: drag and drop from list view to tree view = copy, drag and drop from tree view to tree view = move. You can create an entire hierarchy and even copy the same code under different categories --bart

Should we allow the user to create multiple hierarchies, for example, to test if a categorization works well? --bart

Visualizing codes

Displaying only a set of codes in an interview

  1. TBD

Displaying sentences related to a set of codes

  1. TBD

Using Automated Analyses

Finding code co-occurrences

TBD

TBD

Working in a Team

Sharing a Coding Scheme and Hierarchy

TBD

Sharing a Project with colleagues

TBD

Coding a colleague's interview

TBD

Inter-Coder Agreement

TBD

A new version of Qualyzer comes out

TBD

Updated