< main course page
MAS.961 Social Visualization
March 9, 2003
Mapping the people with whom I have communicated over the past 30 days has been an interesting project. First of all, I wanted to represent my social network as a somewhat organic geography, with a sense of regionality or place in which relationships exist. Therefore, I chose an ego-centric, circular form and then cut the circle according to the various places or settings where I know or interact with people. The size of each region corresponds with the percentage of the number of people in that region.
I placed people I am closer to - or work more closely with - near the center, and people I am acquainted with push to the outer ring. I weighted people within each region-- if they were friends I used to work with who I now consider friends rather than work relationships, they are still weighted towards the work sector. The same goes for riding and even school, although it is problematic in that the sectors are not adjacent to one another. It was still a helpful way to distribute people. I also placed people who knew each other near each other, then drew the amoebas around each group. In the cases where groups overlap, for example, in the friends region, it may be deduced that edw and sb, in the middle ring, introduced me to cp, aw, lmc, etc. In the school region, the groups represent courses I am taking with overlapping of individuals in those courses.
The missing variable in the diagrams is the MOTION variable. I envision each circle as having a vibrating motion according to the amount of communications (frequency, depth) between me and that person, each individual having a number assigned to determine its activity. The amoeba could also have some motion based on the collective rank of that group as compared with the others. The type of motion could also indicate direction or type of communication. I thought this would be more effective- and much less cluttered- than trying to show activity with lines and line-style, given the already busy nature of this landscape.
I explored the following variables:
|Actors/Nodes:||Who do I know?||Individuals I have communicated with over the past 30 days - various media||Circles with initials|
|Groups/Clusters:||How do I know them?||Regional or geographical groups||Sectors on a radial graph, (e.g. friends, family, school). Individuals are weighted towards other sectors as appropriate (good friends I used to work with are weighted towards the work region). Size of sector|
|Who knows who?||Groups of individuals that interact/know each other well||"Amoebas" group people together (e.g. mit class, tl and kbl's group of friends). Individuals who know each other are placed in the regions accordingly. Couples are very close together if I know both well.|
|Relation/ Closeness of Tie:||How well do I know them?||Varying degrees of relationships||Closeness to the center (me) relates to the closeness within that region (e.g. school or friends = know well vs. acquaintance). This tends to relate to the level and intensity (strength) of communication as well.|
|Strength of Communications||How much do I communicate with them?||Quick exchange to in-depth conversation||Motion shows amount of activity. Frequent # of conversations = active. Each circle will vibrate accordingly (still to active) and each amoeba will have an overall amount of motion.|
Other variables I could explore are attributes related to individuals (sex, age, etc.) and attributes of the communications beyond activity, such as type of exchange, direction, roles, etc. These could be mapped with color, shade, and/or size . For example, the length of the conversation could be mapped with color, so that a long conversation could be mapped as gold and a short with a blue, with different levels of saturation representing the depth. A long chit-chat would be a pale gold, whereas a quick, content-heavy exchange would be a deep blue. I also did not assign meaning to the shade of the amoeba clusters, which could certainly be given meaning based on recent activity or frequency of communications. The complexity in that is finding a variable that can be represented by the group vs. the individuals in that group.
Image 1: Distribution of People Across Personal Regions
Image 2: Geographical Features: Individuals Clustered into Subgroups