绘制没有嵌套 for 循环的 Pandas 数据透视表

Question

I have a pivot table written with this code:

piv = pd.pivot_table(df, index = ['trial_id','cluster', 'orientation'], columns = 'spatial_freq', values = 'firing_rate')

and looks like this:

I want to plot the values for each trial/cluster pair. I am able to access these values using .get_level_values() and .xs():

trials = piv.index.get_level_values(0).unique()
trial_1 = piv.xs(trials[0])
clusters = trial_1.index.get_level_values(0).unique()
cluster_1 = trial_1.xs(clusters[0])
cluster_1

out:

I can then run this variable through plt.plot() to get my desired plot. This works for me because I need to label my plots with the trial id and the cluster number. However, with my solution I would need to run a nested for loop and my dataset is very large. Is there a way to achieve what I need with just the original pivot table using Pandas functions such as .apply()?

Thanks in advance!

Answer 1

You just need one loop. You can use groupby to get the sub dataframes:

for name, group in pvt.groupby(level=[0,1]):
    print(f'plotting group {name}') 
    group.plot() # or custom function, use group as a dataframe