# README This project contains supplemental material for the paper **Talk to the Wall: The Role of Speech Interaction in Collaborative Visual Analytics**, accepted at [IEEE VIS 2024][1]. First, the file **exploratory-study.mp4** contains a video presenting the exploratory study. Then, there are three folders. ### Questionnaires The Questionnaires folder contains the following PDF files: - **pre-questionnaire.pdf**: The questionnaire participants filled out before solving the task. - **post-questionnaire.pd**f: The questionnaire participants filled out after solving the task. ### Data The Data folder contains the following CSV files: - **all_interactions_all_participants.csv**: The interactions logged during the study. In the pre-questionnaire sub-folder: - **pre-demographics.csv**: The answers to the pre-questionnaire questions about demographics. - **pre-interaction-experience.csv**: The answers to the pre-questionnaire questions about previous experience with input odalites and interactive systems. - **pre-collaboration-experience.csv**: The answers to the pre-questionnaire questions about previous experience collaborating with others. - **pre-personal-traits-*.csv**: The answers to the pre-questionnaire questions corresponding to the the IPIP-NEO-60 personality assessment instrument. In the post-questionnaire sub-folder: - **post-modality-preferences.csv**: The answers to the post-questionnaire questions about preference of input modality. - **post-collaboration.csv**: The answers to the post-questionnaire questions related to the collaboration experience. - **post-time-distribution.csv**: The answers to the post-questionnaire questions related to the distribution of time among multiple activities. In the figures sub-folder: - **modalities_per_action.csv**: The actions executed by each participant with the used input modality. - **speech_count_with_personality_scores.csv**: The total count of speech commands executed per participant, combined with their scores per personality trait. - **wall-positions.csv**: The positions of the participants in front of the wall, when they executed a speech command or a touch gesture. - **avg_distance_to_wall_with_collab_style.csv**: The average distance between each participant and the wall display, during loose or close collaboration. - **avg_distance_between_participants_with_collab_style.csv**: The average distance between the two participants of each pair, during loose or close collaboration. The answers are anonymized and participants are identified by the IDs going from P1 to P20 (excluding the participants of the pilot studies). They are organized in groups G1 to G10. ### Scripts The Scripts folder contains the following Jupyter notebooks: - **fig3-modality-use-and-preferences.ipynb**: Code to generate Figures 3(a) and 3(b). - **fig4-personality-traits.ipynb**: Code to generate Figure 4. - **fig5-wall-positions.ipynb**: Code to generate Figure 5. ---------- ### How to run the scripts ### The Jupyter notebooks run with Python 3.11.7 and Jupyter Notebook 7.0.8. At the beginning of each notebook, you can find the list of Python libraries required. The scripts recreate Figures 3, 4, and 5 of the paper. 1. Download the csv files in the figures subfolder to a folder called "data". 2. Download the csv file **post-modality-preferences.csv** from the post-questionnaire folder to the data folder. 3. Run each Jupyter Notebook in the parent folder. The pdf versions of the figures will be generated in the parent folder. We finalized Figures 3 and 5 in [Figma][2] for minor fixes (e.g., adding the gray arrows on Figure 3). Figure 6 was created with [RAWGraphs][3] (no coding necessary) and finalized in Figma, using the files **avg_distance_to_wall_with_collab_style.csv** and **avg_distance_between_participants_with_collab_style.csv**. [1]: https://ieeevis.org/year/2024/welcome [2]: https://www.figma.com/ "Figma" [3]: https://www.rawgraphs.io/about "RAWGraphs"