How to look at trend data

Peter Shawhan
Updated November 17, 2002 to reflect new ligo-viewer package

The LIGO data acquisition system produces second-by-second and minute-by-minute summaries (minimum, maximum, mean, and rms values) for essentially every channel as the data is collected. These "trend" data files are very useful for tracking the gross behavior of channels over long periods of time, which would be impractical using the raw data.

The data-acquisition system at each observatory keeps the last few weeks of second-trend data, plus essentially all of the minute-trend data ever generated. Also, all of the trend data is copied to the LIGO tape archive at Caltech in the form of frame files. At present, there is no single application which retrieves and views trend data from both of these storage methods, but there are applications which work in each case.

There are two applications which allow users to view the "on-site" trend data:

The CDS Data Viewer (LIGOtools package "dataviewer") is used extensively in the control rooms in the sites. It should be launched from the DTT main menu window. It only works on Sun computers. Data Viewer can be used on computers outside of the control rooms, even off-site, using ndsproxy as described in this FAQ article.
Vladimir Dergachev's "LIGO Viewer" is a graphical user interface specifically designed for viewing LIGO trend data. It works on Linux, Sun, and probably most other unix computers. It too can be used off-site, retrieving data via the ndsproxy servers at the sites.

If you want to look at old second-trend data that is no longer on disk at one of the observatory sites, then you must retrieve it from the data archive at Caltech. You can do this using the getFrames utility with the appropriate UDN (Universal Dataset Name). The complete list of available UDNs is given on this page. For example, to retrieve some second-trend data from LLO during the E5 engineering run, you could use:

  getFrames -d //ligo/sec/llo/e5 -t 681000000-681001000 -c "L1:ASC-QPDX_DC,L1:ASC-QPDY_DC" -o

This retrieves the requested data into a frame file on your local disk. The best way to look at the data in the file is probably to use Matlab, as described in this FAQ article. Note that for a given raw channel such as L1:ASC-QPDX_DC, there are five channels in the trend data:

  L1:ASC-QPDX_DC.min
  L1:ASC-QPDX_DC.max
  L1:ASC-QPDX_DC.mean
  L1:ASC-QPDX_DC.rms      # root-mean-square WITHOUT first subtracting the mean
  L1:ASC-QPDX_DC.n        # Number of data points used to calculate quantities

getFrames is smart enough to get all of these when you just specify the base name (L1:ASC-QPDX_DC in this case), but you will have to specify an explicit channel name, e.g. L1:ASC-QPDX_DC.mean, to read into Matlab. Note that the ".rms" channel is not what you would probably expect it to be; to get the more usual root-mean-square-deviation-from-mean-value, you would have to do some arithmetic on your own.