indicate_error for x direction #194

Activity

1. lukelbd

   added

   feature

   on Jun 18, 2020
2. lukelbd

   closed this as completedin #258on Jul 11, 2021
3. lukelbd commented on Jul 11, 2021

   

   lukelbd

   on Jul 11, 2021

   Last edited by lukelbd

   Collaborator

   More actions

   Added this as part of a giant refactor. Here's the massive new website example showcasing this:

   import proplot as plot
   import numpy as np
   import pandas as pd
   
   # Sample data
   state = np.random.RandomState(51423)
   data = state.rand(20, 8).cumsum(axis=0).cumsum(axis=1)[:, ::-1]
   data = data + 20 * state.normal(size=(20, 8)) + 30
   data = pd.DataFrame(data, columns=np.arange(0, 16, 2))
   data.columns.name = 'column number'
   data.name = 'variable'
   
   # Loop through "vertical" and "horizontal" versions
   array_vertical = [[1], [2], [3]]
   array_horizontal = [[1, 1], [2, 3], [2, 3]]
   for name, array in zip(('horizontal', 'vertical'), (array_horizontal, array_vertical)):
       # Figure
       fig, axs = plot.subplots(
           array, refaspect=1.5, refwidth=4,
           share=0, hratios=(2, 1, 1)
       )
       axs.format(abc=True, suptitle=f'Indicating {name} error bounds')
   
       # Medians and percentile ranges
       ax = axs[0]
       kw = dict(
           color='light red', legend=True,
           median=True, barpctile=90, boxpctile=True,
           # median=True, barpctile=(5, 95), boxpctile=(25, 75)  # equivalent
       )
       if name == 'horizontal':
           ax.barh(data, **kw)
       else:
           ax.bar(data, **kw)
       ax.format(title='Bar plot')
   
       # Means and standard deviation range
       ax = axs[1]
       kw = dict(
           color='denim', marker='x', markersize=8**2, linewidth=0.8,
           label='mean', shadelabel=True,
           mean=True, shadestd=1,
           # mean=True, shadestd=(-1, 1)  # equivalent
       )
       if name == 'horizontal':
           ax.scatterx(data, legend='b', legend_kw={'ncol': 1}, **kw)
       else:
           ax.scatter(data, legend='ll', **kw)
       ax.format(title='Marker plot')
   
       # User-defined error bars
       ax = axs[2]
       means = data.mean(axis=0)
       means.name = data.name
       shadedata = np.percentile(data, (25, 75), axis=0)  # dark shading
       fadedata = np.percentile(data, (5, 95), axis=0)  # light shading
       method = ax.plotx if name == 'horizontal' else ax.plot
       kw = dict(
           shadedata=shadedata, fadedata=fadedata,
           label='mean', shadelabel='50% CI', fadelabel='90% CI',
           color='ocean blue', barzorder=0, boxmarker=False,
       )
       if name == 'horizontal':
           ax.plotx(means, legend='b', legend_kw={'ncol': 1}, **kw)
       else:
           ax.plot(means, legend='ll', **kw)
       ax.format(title='Line plot')

   
4. lukelbd commented on Jul 11, 2021

   

   lukelbd

   on Jul 11, 2021

   Last edited by lukelbd

   Collaborator

   More actions

   I went about this by adding plotx and scatterx commands (note the trailing x), analogous to fill_betweenx, hlines, and barh, that use the y axis as the dependent variable. They take input arguments as e.g. ax.plotx(y) or ax.plotx(y, x). Had been meaning to do this anyway! The website now showcases these commands under the "1D plotting" section.