The dimensions of the matplotlib plot are determined by the range of the values plotted. How can I specify them separately? With the bokeh backend this would be done with `width`

or `height`

. I had assumed the correct value would be `fig_size`

, but these appears to only take scalar values.

Here’s an example of what’s going wrong:

```
import holoviews as hv
import numpy as np
import pandas as pd
hv.extension("matplotlib")
df = pd.DataFrame({"x": np.random.randn(100000), "y": np.random.randn(100000)}, index=range(100000))
# This looks okay:
display(datashade(hv.Points(df, ["x", "y"])).opts(fig_size=200))
```

```
# This does not:
df["x2"] = df["x"] * 10
display(datashade(hv.Points(df, ["x2", "y"])).opts(fig_size=200))
```