DESIGN ANALYZER
A creativity support tool built on application of visual analytics in parametric design domain.
Research paper submitted for final review in CAADfutures’19. Extended abstract accepted.
- Tasks: HCI Research, UX Research, UX design
- Tools used: Adobe Indesign, Adobe XD, Rhinoceros 3D, Grasshopper
- Project Duration: 3 months
- Association: Dr. Halil Erhan, Computational Design Lab, SIAT, SFU, Canada
Description of Purpose
Current parametric design tools are limited in terms of the cognitive aid they provide for comprehension of large number of parametric alternatives. This poses a choice overload problem for the designers.
Visualization of parametric data pertaining to these design alternatives allows the designer to explore these alternatives in more depth with simultaneous comparison, increasing efficiency of the whole design-decision making process. To test this theory, I have designed a design analytics tool named ‘Design Analyzer’ which primarily enables data visualization of parametric data.
RESEARCH
Problem Domain
The research was set around the study of choice overload problem prevalent in the parametric design domain. With an increasing amount of design “Big- Data” produced rapidly in the process of parametric design, the design decision- making process is made more complex adding to the cognitive load and reducing efficiency.
Design Analytics
System features
A visual analytics tool for parametric design data.
Main Features
- Parametric Analysis
- Repository
- Timeline
- Hierarchical Clusters
Home
- The home screen acts as a dashboard offering a synopsis of latest activity in the system, Timeline of project analysis, Hierarchical clusters and group activity.
- The user can check notifications to see any new messages or activity from collaborators on the projects.
Repository
- The system has a repository which acts as an archive of all the different project alternatives imported into the system from various sources. All the alternatives are stored along with their metadata which is later seen in expanded views when then analyzer is launched.
- The user can add new files at any time by clicking on the add new button as shown.
Launch Analyzer
- The user launches the analyzer by clicking on the “Analyzer” button as shown. From there on, all the alternatives present in the repository are visible in the left hand side segment of the screen. The user can then select some or all alternatives to analyze.
- All the metadata pertaining to individual alternatives is listed below every alternative and automatically stacked into parametric nature based data sets. These data sets can be seen in the right segment of the screen. Several options for visualizations are available to the user. The user can then perform some basic impromptu parametric analytics.
Analyzer in expanded view
- For detailed parametric analysis, the user can switch to an expanded view of the analyzer.
- This view offers more options like saving, pinning and viewing multiple visualizations, sharing them, etc.
- To facilitate simultaneous parametric analysis, I created the option to pin alternatives.
Explore Selection Mode
- For detailed parametric analysis, the user can interact with live visualizations, selecting data points that lie in their preferred parametric zones. The user can then right click and enter the explore selection mode.
- The explore selection mode automatically re-selects the alternatives based on the data points in the selected zone and re-configures the data sets accordingly.
- From here, the user can create new visualizations to explore the selected alternatives deeper.
To experience the system yourself, please click on the interactive prototype link here for your perusal.
Design Process
Conceptualization & Design
This system was built keeping in mind the complexity of thought process which needs to be enabled for its primary users i.e., parametric designers. The nature of tasks being performed are majorly creative, hence, creating a system that supports such creative activities is a monumental challenge of its own. The process of conceptualization and design was done in 3 phases as explained below.
PHASE I
A brief overview of the conceptualization phase with sketches.
PHASE II
A use case diagram showing interactions in the ‘Parametric Analytics’ task.
An activity diagram showing the interaction b/w user & prototype in “preferential analytics” mode.
Holistic class diagram showing inter-relationships at numerous levels and class hierarchy in the system.
PHASE III
ITERATE
Low Fidelity PROTOTYPE
The process of conceptualization and formalization of this prototype involved detailed System Requirement Analysis and End-User studies. The development was done in phases involving rapid iterations. In the end, a low fidelity, interactive prototype was conceptualized and built in Adobe XD.
User Testing
For Evaluation of this prototype, a semi-controlled experiment setting was created. 6 participants with varied levels of expertise in parametric design domain were assigned the same task: to explore 12 different alternatives of a rectangular prism, imported from the Design Gallery. After completing the set tasks, the participants were asked 12 questions based on the Creativity Support Index: CSI (Cherry and Latulipe 2014). The questions asked participants to express their experience using the too in the form of ratings on a scale of 1 to 10. The responses were recorded and computed using CSI method and inferences were derived from this data.
6 PARTICIPANTS
A TASK OF EXPLORING 12 DESIGN ALTERNATIVES
CREATIVITY SUPPORT INDEX (CSI) USED FOR EVALUATION
On completion of prototype exploration, all the participants were asked a set of questions based on CSI and their responses were recorded for further analysis. The CSI enables a researcher to evaluate their tool based on the following factors:
- Collaboration
- Enjoyment
- Exploration
- Expressiveness
- Immersion
- Results Worth Effort
The results obtained upon analysis of the collected data from the 6 participants are shown in tables below.
RESULTS
Inferences based on the evaluation:
- Looking at the cumulative scores, a positive feedback from all users was received.
- The CSI score value was greater than 50 in all cases reflecting on the positivity of feedback obtained.
- The average CSI score is 70 if cumulatively computed for all 6 participants.
- Based on the CSI score for individual factors, Collaboration is the aspect with maximum dissatisfaction among the users, and further improvement.
- The users were most satisfied with the enjoyment factor of the tool from the CSI scores received.
“The Analytical functions of the system are amazing."
Participant 1
“The system is easy to use. I would use it.”
Participant 4
ITERATIONS
Final Design
project insight & tradeoffs
- Deciding on the mode of application and Ui design required instensive research and testing.
- The concept of impromptu analytics as a result of the exploration for screen ratios for spatial allocation of content.
- The amount of direct linkage with Grasshopper was another tradeoff made as a result of the research outcome. Since a clear archival output cannot be received from the software and there is a need for an intermediary acting as a repository. The idea of a repository is based on the basic nature of this intermediary and is intended for further development.
- There was a constant tradeoff between the time constraint and research goals directly affecting the number of issues I could address leading to formation of direct features in the prototype.
- This project made me appreciate the potential and complexity of data and data analytics and how ubiquitous its application can be.
