Ekaterina Obyedkova

Final project for the MIT class 4.566 Advanced Project in Digital Media
Development: February-May, 2013
Instructor: Prof. Takehiko Nagakura

Crowd sourcing experiment demonstration

The video shows an overview of the four experiments and the two questionnaires. This video was created in order to prepare test subjects for the experiment.

Click here to enlarge the video.

Movie info: mp4, 4:09 minutes

Project Overview

Computational technologies have become integral part of design process. Not only computer aided design replaced traditional drawing boards, it made possible new paradigms in design thinking and architecture. Novel design approaches most often exploit new ways of dealing with quantitative information and large datasets. What if, however, computers could have qualitative understandings of space that is very similar to ours? Then not only would computers be able to assist problem solving but also describe what spatial experiences a particular environment creates. In this project I decided to look at such qualitative notions of space as Public and Private. More specifically, at the qualitative correspondence between discrete words and continuous space.

Fig 1. Research hypothesis

Such terms as public and private escape precise definition because they comprise a number of factors such as geometry, lighting, colour, texture, style and culture. In order to study differences and similarities in spatial perception of these abstract terms I have created a synthetic environment, Morphing Space, in which spatial variation is limited to height, width, lighting and colour.

Fig. 1 gives an example of a public space and a private space.

Fig 2. Experiment 1

The first experiment is based on an animation of Morphing Space. User's task is to define which parts of the animation represent public space and which represent private space.

Fig 3. Experiment 2,3 and 4

Experiment 2 is based on a slide show of images. Users are asked to tag all images that show public space. The images are randomly mixed frames from the animation. Experiment 3 is also image based but for every image the user is asked to rate it on the scale of 8. In Experiment 4 the users are given programmatic description of space prior rating every image on the scale of 8.

Fig 4. Data sample and Questionnaires

Test subjects were asked to fill in two questionnaires: personal information(prior the experiment) and subjective feedback(after the experiments).

Fig 5. Data Analysis and Contributions

1) Developed a crowd sourcing experiment, consisting of 4 parts.
2) Demonstrated that there are significant differences in spatial perception.
3) Showed that spatial perception is impacted by verbal instructions.
4) Demonstrated that spatial perception is relative to the most recent experiences.

Further development

1) Further research on design concepts and spatial perception can incorporate other notions, such as open space, narrow space etc.
2) Significant understanding of cultural variation in spatial perception can be achieved by providing the user with a more interactive space modelling environment.
3) A more robust analysis can be performed by increasing the user population as well as performing such statistical tests as multivariable ANOVA and t Test.