Loading presentation...

Present Remotely

Send the link below via email or IM

Copy

Present to your audience

Start remote presentation

  • Invited audience members will follow you as you navigate and present
  • People invited to a presentation do not need a Prezi account
  • This link expires 10 minutes after you close the presentation
  • A maximum of 30 users can follow your presentation
  • Learn more about this feature in our knowledge base article

Do you really want to delete this prezi?

Neither you, nor the coeditors you shared it with will be able to recover it again.

DeleteCancel

Eyeson

No description
by

Josh H

on 22 April 2010

Comments (0)

Please log in to add your comment.

Report abuse

Transcript of Eyeson

Visualizing Groups Network Diagrams Pro:
Groupings Evident
Maintains Individual Level Data Con:
Cluttered and complex
Requires knowledge of individual connections
Assumption:
Similarity implies connection Principal Component Analysis (PCA) Cluster Analysis Our alternative... Eyes on Elis:
Constellation Modeling Our Data: What we know about you. Year Hometown (Income) Major College Extra Curriculars (some) But not who knows who! Can't use Network Diagrams! Hierarchical Cluster Analysis Ugly! But PCA tells us something is going on here! Pro:
Highlights the groupings we're interested in
Mathematically Based
Con
Potentially ambiguous
Limited graphical representations at higher
dimensions
Takes a loooooonnnnnngggg time
Pro:
Useful for discovering underlying order
Highlights groups
Con:
Difficult to visualize >2 components
Question: How to visualize many variables with many nominal responses that split into many groups? Inspired by WeFeelFine and
Cosmic Architecture
Our Project! Each dot is a student at yale, moving in a manner that is based on gravitation and basic principles of cluster analysis, constrained by computing power concerns. Every dot is attracted to every other dot based on their similarity. The importance of each simlarity is user defined through a real-time user-interface (Thanks Alark!) At each iteration, the dot looks at its n nearest neighbors (where n is predefined to optimize running speed) and will move in the direction of the majority of similar dots (accomplished through weighted vector addition.) Eventually, constellations form. As groups become more definite they all take on the same color in order to assist in visual identification. At any time the user can pause the prgram and select one or constellations for a print read-out of the students contained within, including relavant grouping characteristics.
Groupings (specifically that there is a big split between rich newenglenders and everyone else and that there is a big difference between upper and lower classmen)
Statistical Methods
A Work in Progress... Still need to...
Incorperate more intuitive visual cues to convey information about clusters (better use of size, color)
Optimize program for more reasonable run-time
Refine grouping algorithim to handle more nuanced clusters
impliment tool tips...
....and search bar!!! Suggestions?
Full transcript