Loading Data for Analysis

KlustaWin can obtain its source data from a text file or from text data placed on the clipboard.

Source file and clipboard format

Each data item occupies a separate row. The dimensional values of each data item are in tab-separated columns. Each item must have the same number of dimensions, with no missing values. There can be an arbitrary number of text lines at the start of the file. Lines containing characters other than whitespace, the digits 0-9, the decimal point, the plus and minus signs, the letter E or e (for scientific notation) are considered to be text lines and are ignored. Text lines should only occur at the start of the file.

Here is an example of the start of a source file:

File contains the 1st 3 principal components of a set of
spike waveforms recorded extracellularly from the abdominal
superficial 3rd roots of a crayfish.
0.0578           -0.0018           -0.0241
0.0680           -0.0395            0.1050
0.0576           -0.0028           -0.0342
0.0154           -0.0033            0.0100
0.0267           -0.0217            0.0557
0.0610           -0.0024           -0.0341
0.0577           -0.0068           -0.0243
0.0486           -0.0504            0.1099

KlustaWin does some simple parsing to try to make sure that the file contains recognizable data and to catch obvious errors, but it is almost certainly possible to crash the program by loading source files that are incorrectly formatted.

Loading files

There are five ways of loading data into Klustawin

Command line parameter
Start the Klustawin programme with a filename of the appropriate format given as a command line parameter.

Load
Click the Load button in the Source panel, and select the file using the standard Open dialog.

Drag-and-Drop
Drop the file onto the program window using drag-and-drop from Windows Explorer.

Paste
Click the Paste button to insert text data which has been placed on the clipboard with the same format as a source data file.

When the program has loaded the data file, the number of Items in the file and the number of Dimensions for each item are displayed.

NOTE: after you have loaded a data file you can load a partition file to display the results of a previous analysis of that file.

Axes

Up to 3 dimensions of the source data are displayed on the scatter graph. You can choose which dimensions are displayed on the graph using the X, Y and Z spin buttons in the Axes/Dimensions panel.

Prior to clustering the data items are all displayed in a single colour. This colour can be changed by clicking the coloured block under the Clusters drop-down list in the Analysis panel, and selecting a new colour from the dialog.

See Observing 3-D data for more details of how to use the display features.