viewer

Interactive analysis of time-series data.

Interactively analyze time-series data ...

... even with small windows sizes.

Functions

• viewer.ts() ... Viewer for time-series data

Details

Interactive viewer for time-series data. Replaces the older “ui.viewer”. Although it is a user interface utility, it is large enough to make up its own module.

Variable types that can in principle be plotted are:

• np.ndarray
• pd.core.frame.DataFrame
• pd.core.series.Series

Viewer can be used to inspect a single variable, or to select one from the current workspace.

Notable aspects:

• Based on Tkinter, to ensure that it runs on all Python installations.
• Resizable window.
• Keyboard-based interaction.
• Logging of marked events.

viewer.ts(data=None)

Show the given time-series data. In addition to the (obvious) GUI-interactions, the following options are available:

Keyboard interaction:

• f ... forward (+ 1/2 frame)
• n ... next (+ 1 frame)
• b ... back ( -1/2 frame)
• p ... previous (-1 frame)
• z ... zoom (x-frame = 10% of total length)
• a ... all (adjust x- and y-limits)
• x ... exit

Optimized y-scale:

Often one wants to see data symmetrically about the zero-axis. To facilitate this display, adjusting the “Upper Limit” automatically sets the lower limit to the corresponding negative value.

Logging:

When “Log” is activated, right-mouse clicks are indicated with vertical bars, and the corresponding x-values are stored into the users home-directory, in the file “[varName].log”. Since the name of the first value is unknown the first events are stored into “data.log”.

Load:

Pushing the “Load”-button shows you all the plottable variables in your namespace. Plottable variables are:

• ndarrays
• Pandas DataFrames
• Pandas Series

Examples:

To view a single plottable variable:

>>> x = np.random.randn(100,3)
>>> viewer.ts(x)

To select a plottable variable from the workspace

>>> x = np.random.randn(100,3)
>>> t = np.arange(0,10,0.1)
>>> y = np.sin(x)
>>> viewer.ts(locals)