plot pandas DataFrame with condition columns

Question

I have this kind of pandas.DataFrame. "a","b" are conditions when getting "x" and "y".

df = pd.DataFrame([[10,20,0,.1], [10,20,1,.5], [100,200,0,.33], [100,200,1,.11]], columns=["a", "b", "x", "y"])
df

I need to plot line charts of (x,y) colums with respect to the same condition. The expected result plot is:

Of course, this image is manually given by the following code:

pd.DataFrame([[0,.1],[1,.5]]).plot(kind="line", x=0, y=1, style="-", legend=None, title="a: 10, b: 20")
plt.xlabel("x")
plt.ylabel("y")
plt.figure()
pd.DataFrame([[0,.33],[1,.11]]).plot(kind="line", x=0, y=1, style="-", legend=None, title="a: 100, b: 200")
plt.xlabel("x")
plt.ylabel("y")

My question is how to dynamically make plots like above when getting a dataframe including condition columns, x and y.

Update

Column names are fixed. However, values of condition columns is dynamically changed. So, I can not use the values 10, 20, 100, 200.

Update2

If I have the below "filter_with_a_and_b" method, I think the problem solved:

def filter_with_a_and_b(df, a_b):
    # how to implement?

a_b_list = df.drop_duplicates(["a","b"])
new_df_list = filter_with_a_and_b(df, a_b)
for idx, df in enumerate(new_df_list):
    df.plot(title=a_b_list[idx])

what is the source for your conditions and on which columns do you want to apply those conditions? — MaxU - stand with Ukraine, Mar 16 '16 at 13:17
so you want to filter columns `a` and `b` with values of columns `a` and `b`, which would produce the same data frame - is that what you want? ;) — MaxU - stand with Ukraine, Mar 16 '16 at 13:24
do you just want to plot (x, y) for all unique combinations of (a, b)? I.e. if you have three different combinations of (a,b) - that would produce three plots with corresponding (x,y) - do i understand it correctly? — MaxU - stand with Ukraine, Mar 16 '16 at 13:29

MaxU - stand with Ukraine · Accepted Answer · 2017-05-06T18:49:04.747

16

is that what you want?

df.loc[(df.a == 10) & (df.b == 20), ['x','y']].plot(title='a: 10, b: 20')

and now let's do it bit smarter:

cond = {'a': 100, 'b': 200}
df.loc[(df.a == cond['a']) & (df.b == cond['b']), ['x','y']].plot(title='a: {a}, b: {b}'.format(**cond))

or using query():

q = 'a == 100 and b == 200'
df.query(q)[['x','y']].plot(title=q)

UPDATE:

a_b_list = df[['a','b']].drop_duplicates()

[df.loc[(df.a == tup[0]) & (df.b == tup[1]), ['x','y']] \
   .plot(x='x', y='y', kind='line', style='-',title='a: {0[0]}, b: {0[1]}'.format(tup)) \
   .set_ylabel('y')
 for tup in a_b_list.itertuples(index=False)]

edited May 06 '17 at 18:49

answered Mar 16 '16 at 13:06

MaxU - stand with Ukraine

205,989
36
386
419

@MaxU Sorry, I mean condition values should be dynamically changed. – Light Yagmi Mar 16 '16 at 13:13
1

@rkjt50r983, how do you want to change them dynamically? Could you make an example? – MaxU - stand with Ukraine Mar 16 '16 at 13:14

plot pandas DataFrame with condition columns

Update

Update2

1 Answers1

Linked

Related