I am creating a series of boxplots in order to compare different cancer types with each other (based on 5 categories). For plotting I use seaborn/matplotlib. It works fine for most of the cancer types (see image right) however in some the x axis collapses slightly (see image left) or strongly (see image middle)
https://i.imgur.com/dxLR4B4.png
Looking into the code how seaborn plots a box/violin plot https://github.com/mwaskom/seaborn/blob/36964d7ffba3683de2117d25f224f8ebef015298/seaborn/categorical.py (line 961)
violin_data = remove_na(group_data[hue_mask])
I realized that this happens when there are too many nans
Is there any possibility to prevent this collapsing by code only
I do not want to modify my dataframe (replace the nans by zero)
Below you find my code:
boxp_df=pd.read_csv(pf_in,sep="\t",skip_blank_lines=False)
fig, ax = plt.subplots(figsize=(10, 10))
sns.violinplot(data=boxp_df, ax=ax)
plt.xticks(rotation=-45)
plt.ylabel("label")
plt.tight_layout()
plt.savefig(pf_out)
The output is a per cancer type differently sized plot
(depending on if there is any category completely nan)
I am expecting each plot to be in the same width.
Update
trying to use the order parameter as suggested leads to the following output:
https://i.imgur.com/uSm13Qw.png
Maybe this toy example helps ?
|Cat1|Cat2|Cat3|Cat4|Cat5
|3.93| |0.52| |6.01
|3.34| |0.89| |2.89
|3.39| |1.96| |4.63
|1.59| |3.66| |3.75
|2.73| |0.39| |2.87
|0.08| |1.25| |-0.27
Update
Apparently, the problem is not the data but the length of the title
https://github.com/matplotlib/matplotlib/issues/4413
Therefore I would close the question
#Diziet should I delete it or does my issue might help other ones?
Sorry for not including the line below in the code example:
ax.set_title("VERY LONG TITLE", fontsize=20)
It's hard to be sure without data to test it with, but I think you can pass the names of your categories/cancers to the order= parameter. This forces seaborn to use/display those, even if they are empty.
for instance:
tips = sns.load_dataset("tips")
ax = sns.violinplot(x="day", y="total_bill", data=tips, order=['Thur','Fri','Sat','Freedom Day','Sun','Durin\'s Day'])
Related
I have a DataFrame of Heart Disease patients, which has over 300 values. What I have done initially is filter the patients aging over 50. Now I am trying to plot that DF, but running on Google, I found this piece of code that helped me plotting it.
But I am not able to understand the concept of ax = ax here:
fig, ax = plt.subplots()
over_50.plot(x="age",
y="chol",
c="target",
kind="scatter",
---------> ax=ax); <---------
I want to learn the concept behind this little piece of code here. What is it doing at its core?
In this case (a single axes plot) you can do without this parameter.
But there are more complex cases, when you create subplots with
a number of axes objects (a grid).
In this case ax (the second result from plt.subplots()) is an array
of axes objects.
Then, creating each plot, you should specify in which axes this plot
is to be created.
See e.g. https://matplotlib.org/3.1.0/gallery/subplots_axes_and_figures/subplots_demo.html
and find title Stacking subplots in one direction.
It contains such example:
fig, axs = plt.subplots(2)
fig.suptitle('Vertically stacked subplots')
axs[0].plot(x, y)
axs[1].plot(x, -y)
Here:
there is created a figure composed of 2 columns,
in the first axes there is created one line plot, and in the second - another plot.
Alternative form of how to specify axes object in which particular plot
is to be created is just ax parameter, like in our code,
where you can pass one of axes objects from the current figure.
I am plotting a relplot with Seaborn, but getting the legend (and an empty axis plot) printed under the main plot.
Here is how it looks like (in 2 photos, as my screen isn't that big):
Here is the code I used:
fig, axes = plt.subplots(1, 1, figsize=(12, 5))
clean_df['tax_class_at_sale'] = clean_df['tax_class_at_sale'].apply(str)
sns.relplot(x="sale_price_millions", y='gross_sqft_thousands', hue="neighborhood", data=clean_df, ax=axes)
fig.suptitle('Sale Price by Neighborhood', position=(.5,1.05), fontsize=20)
fig.tight_layout()
fig.show()
Does someone has an idea how to fix that, so that the legend (maybe much smaller, but it's not a problem) is printed next to the plot, and the empty axis disappears?
Here is my dataset form (in 2 screenshot, to capture all columns. "sale_price_millions" is the target column)
Since you failed to provide a Minimal, Complete, and Verifiable example, no one can give you a final working answer because we can't reproduce your figure. Nevertheless, you can try specifying the location for placing the legend as following and see if it works as you want
sns.relplot(x="sale_price_millions", y='gross_sqft_thousands', hue="neighborhood", data=clean_df, ax=axes)
plt.legend(loc=(1.05, 0.5))
I'm using pandas to work with a data set and am tring to use a simple line plot with error bars to show the end results. It's all working great except that the plot looks funny.
By default, it will put my 2 data groups at the far left and right of the plot, which obscures the error bar to the point that it's not useful (the error bars in this case are key to intpretation so I want them plainly visible).
Now, I fix that problem by setting xlim to open up some space on either end of the x axis so that the error bars are plainly visible, but then I have an offset from where the x labels are to where the actual x data is.
Here is a simplified example that shows the problem:
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
df6 = pd.DataFrame( [-0.07,0.08] , index = ['A','B'])
df6.plot(kind='line', linewidth=2, yerr = [ [0.1,0.1],[0.1,0.1 ] ], elinewidth=2,ecolor='green')
plt.xlim(-0.2,1.2) # Make some room at ends to see error bars
plt.show()
I tried to include a plot (image) showing the problem but I cannot post images yet, having just joined up and do not have anough points yet to post images.
What I want to know is: How do I shift these labels over one tick to the right?
Thanks in advance.
Well, it turns out I found a solution, which I will jsut post here in case anyone else has this same issue in the future.
Basically, it all seems to work better in the case of a line plot if you just specify both the labels and the ticks in the same place at the same time. At least that was helpful for me. It sort of forces you to keep the length of those two lists the same, which seems to make the assignment between ticks and labels more well behaved (simple 1:1 in this case).
So I coudl fix my problem by including something like this:
plt.xticks([0, 1], ['A','B'] )
right after the xlim statement in code from original question. Now the A and B align perfectly with the place where the data is plotted, not offset from it.
Using above solution it works, but is less good-looking since now the x grid is very coarse (this is purely and aesthetic consideration). I could fix that by using a different xtick statement like:
plt.xticks([-0.2, 0, 0.2, 0.4, 0.6, 0.8, 1.0], ['','A','','','','','B',''])
This gives me nice looking grid and the data where I need it, but of course is very contrived-looking here. In the actual program I'd find a way to make that less clunky.
Hope that is of some help to fellow seekers....
I am making log-log plots for different data sets and need to include the best fit line equation. I know where in the plot I should place the equation, but since the data sets have very different values, I'd like to use relative coordinates in the annotation. (Otherwise, the annotation would move for every data set.)
I am aware of the annotate() function of matplotlib, and I know that I can use textcoords='axes fraction' to enable relative coordinates. When I plot my data on the regular scale, it works. But then I change at least one of the scales to log and the annotation disappears. I get no error message.
Here's my code:
plt.clf()
samplevalues = [100,1000,5000,10^4]
ax = plt.subplot(111)
ax.plot(samplevalues,samplevalues,'o',color='black')
ax.annotate('hi',(0.5,0.5), textcoords='axes fraction')
ax.set_xscale('log')
ax.set_yscale('log')
plt.show()
If I comment out ax.set_xcale('log') and ax.set_ycale('log'), the annotation appears right in the middle of the plot (where it should be). Otherwise, it doesn't appear.
Thanks in advance for your help!
It may really be a bug as pointed out by #tcaswell in the comment but a workaround is to use text() in axis coords:
plt.clf()
samplevalues = [100,1000,5000,10^4]
ax = plt.subplot(111)
ax.loglog(samplevalues,samplevalues,'o',color='black')
ax.text(0.5, 0.5,'hi',transform=ax.transAxes)
plt.show()
Another approach is to use figtext() but that is more cumbersome to use if there are already several plots (panels).
By the way, in the code above, I plotted the data using log-log scale directly. That is, instead of:
ax.plot(samplevalues,samplevalues,'o',color='black')
ax.set_xscale('log')
ax.set_yscale('log')
I did:
ax.loglog(samplevalues,samplevalues,'o',color='black')
How do you add an axis to the outside of another axis, keeping it within the figure as a whole? legend and colorbar both have this capability, but implemented in rather complicated (and for me, hard to reproduce) ways.
You can use the subplots command to achieve this, this can be as simple as py.subplot(2,2,1) where the first two numbers describe the geometry of the plots (2x2) and the third is the current plot number. In general it is better to be explicit as in the following example
import pylab as py
# Make some data
x = py.linspace(0,10,1000)
cos_x = py.cos(x)
sin_x = py.sin(x)
# Initiate a figure, there are other options in addition to figsize
fig = py.figure(figsize=(6,6))
# Plot the first set of data on ax1
ax1 = fig.add_subplot(2,1,1)
ax1.plot(x,sin_x)
# Plot the second set of data on ax2
ax2 = fig.add_subplot(2,1,2)
ax2.plot(x,cos_x)
# This final line can be used to adjust the subplots, if uncommentted it will remove all white space
#fig.subplots_adjust(left=0.13, right=0.9, top=0.9, bottom=0.12,hspace=0.0,wspace=0.0)
Notice that this means things like py.xlabel may not work as expected since you have two axis. Instead you need to specify ax1.set_xlabel("..") this makes the code easier to read.
More examples can be found here.