I would like to create a seaborn scatterplot, using the following dataframe:
df = pd.DataFrame({'A': [1, 2, 3, 4], 'B': [2, 4, 6, 8], 'C':['y', 'y', 'n', 'n'], 'D': [1, 1, 2, 2]})
In my graph A should be the x-variable and B the y-variable. Furthermore I would like to color based on column D. Finally, when C='y' the marker should be open-faced (no facecolor) and when C='n' the marker should have a closed. My original idea was to use the hue and style parameter:
sns.scatterplot(x='A', y='B', data=df, hue='D', style ='C')
However, I did not manage to obtain the graph I am looking for.
One cannot specify entire marker styles (so 'marker' and 'fillstyle' keys in your case) for matplotlib yet. Have a look on the answer to this post.
So the only thing left for you is to use different markers right away and specify them (as list or dictionary)
sns.scatterplot(data=df, x='A', y='B', hue='D', style='C', markers=['o', 's'])
plt.show()
Apparently, it is very hard to even create non-filled markers in seaborn, as this post explains. The only option is to do some matplotlib-seaborn-hybrid thing... So if you accept to plot things twice onto the same axis (one for a filled marker and one for the unfilled markers), you still have to dig yourself into the quirks of seaborn...