Binned diagnostics

Binned diagnostics are added using the add_binned_stat() and add_binned_stat_2d() methods.
Requested statistics will be computed for the field parameter bands defined by the x parameter for mono dimensional binned diagnostics or inside boxes defined by the x and y parameters for two dimensions binned diagnostics.
res_x and res_y parameters allows to define the size of band (1D) or box (2D) in which to compute these statistics.
Add a binned diagnostic computing requested statistics inside bands
defined by values of the x parameter according to its resolution.

Parameters
----------
name
    Name of the diagnostic.
field
    Field on which to compute statistics.
x
    Field to use for the x-axis.
stats
    List of statistics to compute (count, max, mean, median, min, std, var, mad)
res_x
    Min, max and width for the x-axis.
stat_selection
    Selection clip used to invalidate (set to NaN) some bins.
    Valid conditions are:

        * count
        * min
        * max
        * mean
        * median
        * std
        * var
        * mad

    These clips are Python vector clips.
    Examples:

        * count :>= 10 && max :< 100
        * min :> 3
        * median :> 10 && mean :> 9
This kind of diagnostic is plotted as curve plots.
Add a 2D binned diagnostic computing requested statistics inside
boxes defined by values of the x and y parameter according their
respective resolutions.

Binned 2d data and plots can be accessed or created using special keywords:

 * plot="box" (default): color mesh representation, on an x-axis/y-axis grid.
 * plot="curve":

    * axis="x": along x-axis representation of each y-field bin
    * axis="y": along y-axis representation of each x-field bin

 * plot="3d": 3d color mesh representation, on an x-axis/y-axis/z-axis 3d grid.
 * plot="box3d": 3d bins surfaces representation,
   on an x-axis/y-axis/z-axis 3d grid.

Parameters
----------
name
    Name of the diagnostic.
field
    Field on which to compute statistics.
x
    Field to use for the x-axis.
y
    Field to use for the y-axis.
res_x
    Min, max and width for the x-axis.
res_y
    Min, max and width for the y-axis.
stats
    List of statistics to compute (count, max, mean, median, min, std, var, mad)
stat_selection
    Selection clip used to invalidate some bins.
    Valid conditions are:

        * count
        * min
        * max
        * mean
        * median
        * std
        * var
        * mad

    These clips are Python vector clips.
    Examples:

        * count :>= 10 && max :< 100
        * min :> 3
        * median :> 10 && mean :> 9
This kind of diagnostics is plotted as pseudo-color plots with a non-regular rectangular grid or as a set of curves.

Diagnostic setting

Setting 1D binning

In the following example we are setting a 1D binned diagnostic computing the mean of a user defined range_std_ku field inside bands defined by values of the swh field between 0 and 8 using a 0.05 step.
from casys import Field

range_std_ku = Field(
    name="range_std_ku",
    source="IIF(FLAG_VAL.ALTI==0, RANGE_STD.ALTI, DV)",
    unit="m",
)
swh = Field(
    name="swh",
    source="IIF(FLAG_VAL.ALTI==0, SWH.ALTI, DV)",
    unit="m",
)

ad.add_binned_stat(
    name="Ku-band range std fct swh",
    field=range_std_ku,
    x=swh,
    res_x=(0, 8, 0.05),
    stats=["mean"],
)

ad.compute()

Setting 2D binning

In the following example we are setting a 2D binned diagnostic computing the median of a user defined range_std_ku field inside boxes defined by values of the swh field between 0 and 8 with a 0.05 step and the var_wind field between 0 and 25 with a 0.5 step.
A second 2D binned diagnostic is computed with a lower y-resolution (step of 2.5).
var_wind = Field(
    name="wind",
    source="IIF(FLAG_VAL.ALTI==0, WIND_SPEED.ALTI, DV)",
    unit="m/s",
)

ad.add_binned_stat_2d(
    name="Ku-band range std fct (swh, wind_speed)",
    field=range_std_ku,
    x=swh,
    y=var_wind,
    res_x=(0, 8, 0.05),
    res_y=(0, 25, 0.5),
    stats=["median"],
)

ad.add_binned_stat_2d(
    name="Ku-band range std fct (swh, wind_speed) [2]",
    field=range_std_ku,
    x=swh,
    y=var_wind,
    res_x=(0, 8, 0.05),
    res_y=(0, 25, 2.5),
    stats=["median"],
)

ad.compute()

Diagnostic plotting

Plotting 1D binning

1D binned diagnostics are plotted as curve plots.
CasysPlot requires a stat parameter to be given in order to know which statistic to display.
from casys import CasysPlot

plot = CasysPlot(
    data=ad,
    data_name='Ku-band range std fct swh',
    stat="mean",
)

plot.show()
../_images/binned_stat_5_0.png

Plotting 2D binning

2 types of plot are available for the 2D binned diagnostics:

  • pseudo-color plots (default)

  • set of curves representing the different y-field bins.

CasysPlot requires a stat parameter to be given in order to know which statistic to display and an optional plot parameter indicating the type of plot to use.

Pseudo-color plots

2D binned diagnostics can be plotted as pseudo-color plots with a non-regular rectangular grid by providing the “box” value to the plot parameter of the CasysPlot object.

Note

This is the default value and not providing the plot parameter has the same effect.

Low resolution binning

plot = CasysPlot(
    data=ad,
    data_name='Ku-band range std fct (swh, wind_speed) [2]',
    stat="median",
    plot="box",
)

plot.show()
../_images/binned_stat_6_0.png

High resolution binning

plot = CasysPlot(
    data=ad,
    data_name="Ku-band range std fct (swh, wind_speed)",
    stat="median",
    plot="box",
)

plot.show()
../_images/binned_stat_7_0.png

Set of curves

2D binned diagnostics can be plotted as a set of curves with each curve representing a bin of the y field.
This type of plot is created by providing the “curve” value to the plot parameter of the CasysPlot object.

Low resolution binning

plot = CasysPlot(
    data=ad,
    data_name='Ku-band range std fct (swh, wind_speed) [2]',
    stat="median",
    plot="curve",
)

plot.show()
../_images/binned_stat_8_0.png

Warning

This type of plot might become difficult to read if the resolution of the y-field is too high and creates a large number of bins/curves.

High resolution binning

plot = CasysPlot(
    data=ad,
    data_name="Ku-band range std fct (swh, wind_speed)",
    stat="median",
    plot="curve",
)

plot.show()
../_images/binned_stat_9_0.png

3D plot

2D binned diagnostics can be plotted as a 3d surface.
This type of plot is created by providing the “3d” value to the plot parameter of the CasysPlot object.
plot = CasysPlot(
    data=ad,
    data_name="Ku-band range std fct (swh, wind_speed)",
    stat="median",
    plot="3d",
)

plot.show()
../_images/binned_stat_10_0.png
Another 3D plot option is available for 2D binned diagnostics, allowing to display each bin as a separate surfaces, with box3d.
plot = CasysPlot(
    data=ad,
    data_name="Ku-band range std fct (swh, wind_speed)",
    stat="median",
    plot="box3d",
)

plot.show()
../_images/binned_stat_11_0.png