Полагаю, что-то подобное будет работать.
f, a = plt.subplots(1, 1)
# Only create one subplot as all the data is meant to go on one plot only
df.xs('A').plot(ax=a, marker='D', markersize=6, markerfacecolor='red',
markeredgecolor='red')
df.xs('B').plot(ax=a, marker='D', markersize=6, markerfacecolor='green',
markeredgecolor='green')
df.xs('C').plot(ax=a, marker='D', markersize=6, markerfacecolor='blue',
markeredgecolor='blue')