Comment

Comment by Tom Wolfer on August 31, 2010 at 9:47am

Yes, deltas would work. When you say first-order differences, etc....do you mean developing a seasonal factor and applying it to a given period, or, what do you mean?

Comment by Dominic Pouzin on August 30, 2010 at 5:24pm

Hi Tom, we don't support auto-smoothing of the data prior to analysis (only smoothing of data / results after analysis) which would be an interesting addition. I guess you could feed it first-order differences, second-order differences, or deltas taken from a month / a year ago. The algorithm uses neural networks, so it's pretty much a black box (ex: it doesn't differentiate between capturing seasonal variations vs. more general trends).

Comment by Tom Wolfer on August 30, 2010 at 1:45pm

Hey Dominic, the data looks cyclical in nature. Is there any way, in this software, to overlay a seasonality component and smooth the data? What if, for example, there were certain events (eg. in marketing, these would likely be campaigns) that occur at various timepoints, and I wanted to get a better sense of what impact each was having on the data fluctuations without the noise of seasonality interfering?

Comment by Vincent Granville on August 11, 2010 at 11:52am

It must be a false positive. I've checked more thoroughly, but I could not find and risk or threat.

Comment by Jerry Nowatzke on August 11, 2010 at 11:35am

whne i clicked on the image above Google identified 1 identity threat and rated the site unsafe

Comment by Vincent Granville on August 11, 2010 at 11:11am

Which site? The hosting site (www.data-applied.com) seems safe according to siteadvisor.com, and after having performed myself a few basic checks.

Comment by Jerry Nowatzke on August 11, 2010 at 10:59am

The site has identity/virus threats and is unsafe to visit

Comment by Jon Swift on July 8, 2010 at 1:06pm

Very interesting, would love to hear what software this is.

Comment by Indrajit Sen Gupta on June 2, 2010 at 10:15pm

Looks interesting. What software is this?