ENH: Scatter plots of one variable vs another #2277
Conversation
|
Could not review pull request. It may be too large, or contain no reviewable changes. |
|
Could not review pull request. It may be too large, or contain no reviewable changes. |
xarray/plot/facetgrid.py
Outdated
| @@ -280,7 +280,8 @@ def map_dataarray_line(self, x=None, y=None, hue=None, **kwargs): | |||
| self : FacetGrid object | |||
|
|
|||
| """ | |||
| from .plot import line, _infer_line_data | |||
| from .plot import (_infer_line_data, _infer_scatter_data, | |||
| line, dataset_scatter) | |||
There was a problem hiding this comment.
E128 continuation line under-indented for visual indent
xarray/plot/plot.py
Outdated
|
|
||
| def dataset_scatter(ds, x=None, y=None, hue=None, col=None, row=None, | ||
| col_wrap=None, sharex=True, sharey=True, aspect=None, | ||
| size=None, subplot_kws=None, add_legend=True, **kwargs): |
There was a problem hiding this comment.
E241 multiple spaces after ','
xarray/plot/plot.py
Outdated
| if size is None: | ||
| size = 3 | ||
| elif figsize is not None: | ||
| raise ValueError('cannot provide both `figsize` and `size` arguments') |
There was a problem hiding this comment.
E501 line too long (82 > 79 characters)
xarray/plot/plot.py
Outdated
| g = FacetGrid(data=ds, col=col, row=row, col_wrap=col_wrap, | ||
| sharex=sharex, sharey=sharey, figsize=figsize, | ||
| aspect=aspect, size=size, subplot_kws=subplot_kws) | ||
| return g.map_dataarray_line(x=x, y=y, hue=hue, plotfunc=dataset_scatter, **kwargs) |
There was a problem hiding this comment.
E501 line too long (90 > 79 characters)
xarray/plot/plot.py
Outdated
| aspect=aspect, size=size, subplot_kws=subplot_kws) | ||
| return g.map_dataarray_line(x=x, y=y, hue=hue, plotfunc=dataset_scatter, **kwargs) | ||
|
|
||
| xplt, yplt, hueplt, xlabel, ylabel, huelabel = _infer_scatter_data(ds, x, y, hue) |
There was a problem hiding this comment.
E501 line too long (85 > 79 characters)
xarray/plot/facetgrid.py
Outdated
| _, _, hueplt, xlabel, ylabel, huelabel = _infer_line_data( | ||
| darray=self.data.loc[self.name_dicts.flat[0]], | ||
| x=x, y=y, hue=hue) | ||
| elif plotfunc == dataset_scatter: |
There was a problem hiding this comment.
I think this bit should be in a separate map_dataset function that can be reused as the Dataset plotting API becomes more complete.
There was a problem hiding this comment.
That's easy to do. thanks.
xarray/plot/facetgrid.py
Outdated
| _, _, hueplt, xlabel, ylabel, huelabel = _infer_line_data( | ||
| darray=self.data.loc[self.name_dicts.flat[0]], | ||
| x=x, y=y, hue=hue) | ||
| darray=self.data.loc[self.name_dicts.flat[0]], |
There was a problem hiding this comment.
E126 continuation line over-indented for hanging indent
xarray/plot/facetgrid.py
Outdated
| self._mappables.append(mappable) | ||
|
|
||
| data = _infer_scatter_data( | ||
| ds=self.data.loc[self.name_dicts.flat[0]], |
There was a problem hiding this comment.
E126 continuation line over-indented for hanging indent
|
This seems like a very cool and useful feature! A few comments:
Happy to provide a more detailed review once the tests are implemented. |
xarray/tests/test_plot.py
Outdated
| class TestScatterPlots(PlotTestCase): | ||
| def setUp(self): | ||
| das = [DataArray(np.random.randn(3, 3, 4, 4), | ||
| dims=['x', 'row', 'col', 'hue'], |
There was a problem hiding this comment.
E127 continuation line over-indented for visual indent
I disagree here. In many cases, we already follow the naming conventions from Seaborn instead, which uses meaningful names. c and s are pretty meaningless. |
|
It is possibly worth taking a look at the recent (not yet released)
I guess I can see the virtue in sticking with matplotlib's old |
|
I don't have an opinion about naming variables and would be happy with whatever decision y'all make. For the code -- I added tests and changed the logic a bit. Following @dcherian's suggestion, now the default behavior is no longer coloring hues with discrete values (legend) but rather with a continuous scale (colorbar). It does make actually more sense and I think it should also be the default behavior for regular line plots. This is the API now: A = xr.DataArray(np.zeros([3, 20, 4, 4]), dims=[ 'x', 'y', 'z', 'w'],
coords=[np.sort(np.random.randn(k)) for k in [3,20,4,4]])
ds=xr.Dataset({'A': A.x+A.y+A.z+A.w,
'B': -0.2/A.x-2.3*A.y-np.abs(A.z)**0.123+A.w**2})
ds.A.attrs['units'] = 'Aunits'
ds.B.attrs['units'] = 'Bunits'
ds.z.attrs['units'] = 'Zunits'
ds.plot.scatter(x='A', y='B')
Specifying ds.plot.scatter(x='A',y='B', hue='z')
ds['z']= ['who', 'let','dog','out']
ds.plot.scatter(x='A',y='B', hue='z')
If you want a discrete legend even for numeric ds.plot.scatter(x='A',y='B', hue='w', discrete_legend=True)
I am a bit bothered by the fact that this is not only a different coloring method, it's a very different style altogether (under the hood using Of course, faceting works as you think it should: ds.plot.scatter(x='A',y='B', hue='z',col='x')
ds.plot.scatter(x='A',y='B', hue='w',col='x', col_wrap=2)
|
|
This is looking really nice. Coincidentally, the new version of Seaborn was released today, and has a whole new doc section on "relational plots": http://seaborn.pydata.org/tutorial/relational.html#relational-tutorial It's probably worth a look over to see if it has good ideas worth stealing, or if we want to make intentional deviations from its behavior in xarray. |
* master: (68 commits) enable sphinx.ext.napoleon (pydata#3180) remove type annotations from autodoc method signatures (pydata#3179) Fix regression: IndexVariable.copy(deep=True) casts dtype=U to object (pydata#3095) Fix distributed.Client.compute applied to DataArray (pydata#3173) More annotations in Dataset (pydata#3112) Hotfix for case of combining identical non-monotonic coords (pydata#3151) changed url for rasterio network test (pydata#3162) to_zarr(append_dim='dim0') doesn't need mode='a' (pydata#3123) BUG: fix+test groupby on empty DataArray raises StopIteration (pydata#3156) Temporarily remove pynio from py36 CI build (pydata#3157) missing 'about' field (pydata#3146) Fix h5py version printing (pydata#3145) Remove the matplotlib=3.0 constraint from py36.yml (pydata#3143) disable codecov comments (pydata#3140) Merge broadcast_like docstrings, analyze implementation problem (pydata#3130) Update whats-new for pydata#3125 and pydata#2334 (pydata#3135) Fix tests on big-endian systems (pydata#3125) XFAIL tests failing on ARM (pydata#2334) Add broadcast_like. (pydata#3086) Better docs and errors about expand_dims() view (pydata#3114) ...
|
I don't know what to do about this test failure: |
|
Yay, tests pass. I'll merge in a few days (cc @pydata/xarray, @yohai ) |
|
Thanks for this
…On Mon, Aug 5, 2019 at 12:04 PM Deepak Cherian ***@***.***> wrote:
Yay, tests pass. I'll merge in a few days (cc @pydata/xarray
<https://github.com/orgs/pydata/teams/xarray>, @yohai
<https://github.com/yohai> )
—
You are receiving this because you were mentioned.
Reply to this email directly, view it on GitHub
<#2277?email_source=notifications&email_token=ABPA5PNDIYQDJME4CUJREDDQDBFQNA5CNFSM4FJJSKRKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOD3SJCVA#issuecomment-518295892>,
or mute the thread
<https://github.com/notifications/unsubscribe-auth/ABPA5PMUANDUGDQKZHTIPYDQDBFQNANCNFSM4FJJSKRA>
.
--
Yohai Bar Sinai
Post Doctoral Fellow
John A. Paulson School of Engineering and Applied Sciences
Harvard University
|
* master: pyupgrade one-off run (pydata#3190) mfdataset, concat now support the 'join' kwarg. (pydata#3102) reduce the size of example dataset in dask docs (pydata#3187) add climpred to related-projects (pydata#3188) bump rasterio to 1.0.24 in doc building environment (pydata#3186) More annotations (pydata#3177) Support for __array_function__ implementers (sparse arrays) [WIP] (pydata#3117) Internal clean-up of isnull() to avoid relying on pandas (pydata#3132) Call darray.compute() in plot() (pydata#3183) BUG: fix + test open_mfdataset fails on variable attributes with list… (pydata#3181)
|
Merging. |
|
Thanks @dcherian ! |
|
Yeah! It owes a lot to your hard work too. Hopefully people find this useful. |
* commit 'f172c673': ENH: Scatter plots of one variable vs another (pydata#2277) Escape code markup (pydata#3189)
whats-new.rstfor all changes andapi.rstfor new APIsize?Say you have two variables in a
Datasetand you want to make a scatter plot of one vs the other, possibly using different hues and/or faceting. This is useful if you want to inspect the data to see whether two variables have some underlying relationships between them that you might have missed. It's something that I found myself manually writing the code for quite a few times, so I thought it would be better to have it as a feature. I'm not sure if this is actually useful for other people, but I have the feeling that it probably is.First, set up dataset with two variables:
Now, we can plot all values of
Avs all values ofB:What a mess. Wouldn't it be nice if you could color each point according to the value of some coordinate, say
w?Huh! There seems to be some underlying structure there. Can we also facet over a different coordinate?
or two coordinates?
The logic is that dimensions that are not faceted/hue are just stacked using
xr.stackand plotted. Only variables that have exactly the same dimensions are allowed.Regarding implementation -- I am certainly not sure about the API and I probably haven't thought about edge cases with missing data or nans or whatnot, so any input would be welcome. Also, there might be a simpler implementation by first using
to_arrayand then using existing line plot functions, but I couldn't find it.