CEB Data Visualization, 2008
Dynamically generated data visualization



The visualization consists of 3 dynamically generated animations based on statistical analysis and aggregation of data provided by the corporation at a regular rate. The data consists of corporate members' communication either making requests for information, or else downloading tools. queries and request for information.

Top screen will feature the intersection of industries (Metasectors) and practices through aggregated summary of searchterms over time

Middle screen will map members' activities based on their geographic locations

Bottom screen will show interrelationships between practices/programs and content downloads.

 



I: Cross-Function Multivariate Stream [click for animation]


Conditions:
• Make visible that search interests span across functions
• Using keyword search terms, show the top search terms in a given timeframe
• Denote the different company practices by color and show the programs by name

Data Used:
• CreateDate, SearchString, ProgramName, MetaSector

Procedure:
• Screens vertically subdivided into all 39MetaSectors
• Horizontal subdivision into 8 practices, labeled and color coded
• Searchstrings are positioned as small dots at the junctions of Program and MetaSector
• They come in bright then fade to grey
• Searchstrings move to new location and get larger based on popularity use
• Movement uses acceleration/velocity envelope




II: GeoMap [click for animation]


Conditions:
• Visually define global reach, show spatially where members are based
• Show color-coded what program or industry they belong to
• Show based on time how active they are

Data Used:
• CreateDate, SearchString, ProgramName or industry, city

Procedure:
• Key global cities are plotted to a gridlike reformatted global map
• Main cities are charted in grey with labels
• Members actions are charted by red marks getting larger with increased activity




III: Rhizome Links [click for animation]


Conditions:
• Show what products / tasks / tools are being accessed (used) by members in a specified

Data Used:
• CreateDate, Practices, ProgramNames, ContentType

Procedure:
• Position the 8 practices dsitributed through the screen space in their color
• Draw satellite notes (programNames) as roots emerging from the practices
• Follow with ProgramNames labels
• Connect ProgramNames across practices by ContentType which will result in an acrobatic swinging of the various programs as they try to line up through the common ContentType words




Data


CreatedDate The time-date that the user activity occurred on the website
SearchString The keywords entered into the search engine on the website
Practice Each member belongs to one of 8 practices
ProgramName The specific research website the user is logged into (46)
User ID Allows for linking users' multiple actions
TitleType Standard level of titles to classify the seniority of users
FK_MetaSector Industries (39)
FK_ContentType The type of content the user downloaded or launched (10)
PostalCode The postal code of the user
FK_StateOrProvince
FK_Country
City


The data is retrieved every 5 minutes



Visualization Conditions


• Visualizations are to be generated from incoming data updated at a rate to be determined
• Visualizations to evolve out of the specific nature of the data
• The intent is to "reveal" or "make visible" significant events through the processing of the data
• Visualizations are to be active and time-base events. Their staging involve visual syntax and visual communication techniques such as drammaturgy, narrative flow, sequencing, spatial and formal interactions
• Visualizations will require a fine balance between expressive form and information revealing



Visualization Methods


• In the 2D space of the screen, the data can be organized according to size, location (x,y), color coding, proximity, etc.
• Time-based visualization allows for narrative development: Action and movement are key components in creating meaning.
• This can occur through the data moving across the screen, data building up from a blank screen, a simple form evolves into a complex form
• The relationship of the data can be determined based on algorithmic modeling, such as Kohonen SOM, bin packing, flocking algorithms, biological simulation models, cellular automata, spanning trees, formal distortions, etc.



Technical Setup

The technical system consists of a server receiving data which will then process the data and feed 3 computers each dedicated to a specific visualization. Each computer will display its data on 3 LCD screens horizontally lined up edge-to-edge. The screens are NEC LCD4020-2-AV panels with dimensions of WxDxH) 36.2 in x 5.5 in x 21 in. Image Aspect Ratio is 16:9 and resolution: 1360X768 each for a total image width of 4080 x 768 pixels.