[go: up one dir, main page]
More Web Proxy on the site http://driver.im/
Divvy View source at GitHub

Divvy is a tool for exploratory data analysis with unsupervised machine learning.

Use Divvy to better understand your scientific data.

Check out a video tour of Divvy.

(To view in HD, watch at vimeo.com.)

Just the basics please.

Divvy is a 64-bit Mac OS X 10.6 or 10.7 (Snow Leopard or Lion) application for performing unsupervised machine learning and visualization. We focus on the clustering (separating data into groups) and dimensionality reduction (finding low dimensional structure in high dimensional data) subfields of machine learning. For visualization we provide support for both the whole dataset (e.g. a scatter plot) and points (e.g. transforming a particular point into an image).

Contact us.

Please direct any questions, comments, feature requests, support emails, &c. to josh@cogsci.ucsd.edu.

Free and open.

The NSF funded this work so that you don't have to. Go get our MIT-licensed code at GitHub. Hack it; improve it; run with it.

Endlessly extensible.

Every clusterer, reducer, point visualizer and dataset visualizer in Divvy is a plugin. We've provided a few big ones (K-means, PCA, scatter plot, &c.) and we're hoping the community will use our plugin protocol to build many more. Each plugin defines its own UI, so your algorithm can look and behave the way that you want it to without top-down constraints.

Have lots of cores?

Divvy is both task and data parallel. No longer will you be waiting for one algorithm to complete before you start another. Start as many as you want and keep using the UI. Only started one? With data parallelism we'll still push your new MacBook Pro to 800% CPU utilization.

Part of your workflow.

Export your clusterings and reductions to .csv and your visualizations to .png. Use your R or Matlab data with our R and Matlab to Divvy export tools.

Tutorial 1: Loading Data

  • Export your dataset from Matlab or R to Divvy.
  • Open a dataset with Divvy.
  • Export Divvy clusterings, reductions and visualizations back to .csv and .png.
A list of 3rd party Divvy plugins.

Joshua M. Lewis, Laurens van der Maaten, Virginia de Sa (2013). Divvy: Fast and Intuitive Exploratory Data Analysis. Journal of Machine Learning Research 14(Oct):3159-3163.

Joshua M. Lewis, Laurens van der Maaten, Virginia de Sa (2012). A Behavioral Investigation of Dimensionality Reduction. N. Miyake, D. Peebles, & R. P. Cooper (Eds.), Proceedings of the 34th Annual Conference of the Cognitive Science Society.

Joshua M. Lewis, Virginia de Sa (2012). Learning Cluster Analysis through Experience. N. Miyake, D. Peebles, & R. P. Cooper (Eds.), Proceedings of the 34th Annual Conference of the Cognitive Science Society.

Joshua M. Lewis, Margareta Ackerman, Virginia de Sa (2012). Cluster Evaluation and Formal Quality Measures. N. Miyake, D. Peebles, & R. P. Cooper (Eds.), Proceedings of the 34th Annual Conference of the Cognitive Science Society.

We build Divvy.

Thanks guys!

  • Source control from GitHub.
  • Styles and layout from Twitter Bootstrap.
  • This material is based upon work supported by the National Science Foundation under Grant No. 0963071. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation.