1

I have an xarray Dataset in python. I would like to plot two dependent (data) variables against each other using the hvplot library. Here's a simple example dataset:

import numpy as np
import xarray as xr
import hvplot.xarray

# Create the dataset
time = np.linspace(0,10)
x = time*2 + 1
y = time*3 - 1

ds = xr.Dataset()
ds.coords['time'] = time
ds['x'] = (['time'],x)
ds['y'] = (['time'],y)

# Output
<xarray.Dataset>
Dimensions:  (time: 50)
Coordinates:
  * time     (time) float64 0.0 0.2041 0.4082 0.6122 ... 9.388 9.592 9.796 10.0
Data variables:
    x        (time) float64 1.0 1.408 1.816 2.224 ... 19.78 20.18 20.59 21.0
    y        (time) float64 -1.0 -0.3878 0.2245 0.8367 ... 27.78 28.39 29.0

It's easy to plot x or y against time just by

ds.x.hvplot()

However what I want is to plot x against y. I would have expected this to work:

ds.hvplot(x='x',y='y')

but this only plots one point at a time, with a slider for the 'time' coordinate. xarray has a plot function which does plot as expected using matplotlib.

ds.plot.scatter(x='x',y='y')

Is there any way to reproduce this with hvplot?

Redlegjed
  • 187
  • 2
  • 11

3 Answers3

2

Another possibility:

ds.y.assign_coords(x=ds.x).hvplot.scatter(x="x", y="y")
Johannes Ringler
  • 36
  • 2
1

Don't know too much about xarray, but the following 2 approaches both work, but honestly I hope someone else comes with a nicer solution:

ds.reset_index(dims_or_levels='time').hvplot.scatter(x='x', y='y')

Or:

ds.hvplot.scatter(x='x', y='y', color='blue').overlay()
Sander van den Oord
  • 10,986
  • 5
  • 51
  • 96
0

Probably the simplest way to plot two data variables against each other is to convert the dataset into a pandas DataFrame. Then plot using hvplot

import hvplot.pandas # to add .hvplot to DataFrames

# Convert Dataset to DataFrame, then select x,y
ds.to_dataframe().hvplot(x='x',y='y')

If the dataset has other variables the plots can be grouped using the 'by' argument

ds.to_dataframe().hvplot(x='x',y='y',by=another_variable)
Redlegjed
  • 187
  • 2
  • 11