Quick and dirty compare¶

Sometimes all your need is a simple comparison of two time series. The modelskill.match() method does just that.

In [2]:

import mikeio 
import modelskill as ms

The model¶

Can be either a dfs0 or a DataFrame.

In [3]:

fn_mod = '../tests/testdata/SW/ts_storm_4.dfs0'
df_mod = mikeio.read(fn_mod, items=0).to_dataframe()  # select 1 item

The observation¶

Can be either a dfs0, a DataFrame or a PointObservation object.

In [4]:

fn_obs = '../tests/testdata/SW/eur_Hm0.dfs0'

match()¶

The match() method will return an object that can be used for scatter plots, skill assessment, time series plots etc.

If the observation or model has more than 1 item, the obs_item or mod_item can be given to specify the relevant item.

In [5]:

cmp = ms.match(fn_obs, df_mod)   # both with 1 item only

C:\Working\Coding\Repos\modelskill\modelskill\obs.py:79: UserWarning: Could not guess geometry type from data or args, assuming POINT geometry. Use PointObservation or TrackObservation to be explicit.
  warnings.warn(

In [6]:

cmp.plot.timeseries();

In [7]:

cmp.skill()

Out[7]:

	n	bias	rmse	urmse	mae	cc	si	r2
observation
eur_Hm0	67	0.052239	0.22824	0.222181	0.174851	0.968321	0.085898	0.929767

In [8]:

cmp.plot.scatter();

By default the regression line is fitted based on all the data. In order to fit do a weighted regression based on the quantiles, use the fit_to_quantiles option.

In [9]:

cmp.plot.scatter(fit_to_quantiles=True);

An example including the background data and skill table using the show_points use the skill_table options:

In [10]:

ms.settings.load_style(name='MOOD')

In [11]:

cmp.plot.scatter(fit_to_quantiles=True, show_points=True, skill_table=True, show_density=True);

Display name of the metrics:
Sometimes you like to modify the display name of the metrics. This can be done by passing a dictionary with the metric name as key and the display name as value to the display_names parameter.

In [19]:

@ms.metrics.metric(display_name="MEAN$_{Model}$", has_units="m")
def mean_mod(obs, mod):
    return mod.mean()

@ms.metrics.metric(display_name="MEAN$_{Obs}$", has_units="m")
def mean_obs(obs, mod):
    return obs.mean()

ms.metrics.add_metric(mean_mod)
ms.metrics.add_metric(mean_obs)

ms.metrics.pr.display_name = '$P_{ratio}$'
cmp.plot.scatter(skill_table=["pr",mean_mod, mean_obs]);

Formatting the metrics:
In this example we modify bias to also add the perentage of the bias relative to the mean of the reference series.

In [ ]:

from modelskill.metrics import bias as mybias

def bias(obs, mod):
    avg = obs.mean()
    bias = mybias(obs, mod)
    sign = "+" if bias > 0 else "-"
    return f"{bias:.2f} ({sign}{bias/avg*100:.1f}%)"

cmp.plot.scatter(skill_table=bias);

In [ ]:

cmp.plot.taylor(normalize_std=True)

Out[ ]: