Subscribe to Dr. Granville's Weekly Digest

Heat Map To Visually Detect Spam Across Networks

Heat Map To Visually Detect Spam Across Networks

X-axis is time, Y-axis is first byte of IP address. The value in each cell represents the ratio between "observed transaction count" and "expected transaction count". Red cells show abnormal spikes in activity, and are suspicious.

Rating:
  • Currently 5/5 stars.

Views: 265

Comment

You need to be a member of AnalyticBridge to add comments!

Join AnalyticBridge

Comment by Gertjan de Lange on September 8, 2008 at 3:25pm
You can also use 'treemaps'. For example with the tool TreeMaps developed by Uni Maryland. No developments are done at this point (as far as I know), but the tool is freely downloadable for testing (http://www.cs.umd.edu/hcil/treemap). Upload a column based data file and you can slice and dice all you want and get these type of trends. Pretty cool I think.
A nice history about these objects can be found at http://www.cs.umd.edu/hcil/treemap-history.
Comment by Mehran Gomari on September 7, 2008 at 11:38pm
You can also write custom S codes to produce similar graphs
Comment by Vincent Granville on September 7, 2008 at 10:47pm
I was told that Tableau (a data mining vizualization software) offers similar features.
Comment by Vincent Granville on August 26, 2008 at 11:12am
The "expected count" for cell xy was computed using as the product of x- and y- marginal counts divided by total count.
Comment by Vincent Granville on August 26, 2008 at 11:09am
It was produced with the new version of Excel (.xlsx), with conditional formatting of the cells based on cell value. Then I selected a range of cells, cut and paste from Excel into Paintbrush, and saved it as an image (png format, a format that I strongly recommend for images).

© 2014   AnalyticBridge.com is a subsidiary and dedicated channel of Data Science Central LLC

Badges  |  Report an Issue  |  Terms of Service