Iterating over columns in data frame by skipping first column and drawing multiple plots

Question

I have a data frame as following, df.head()

     ID AS_FP   AC_FP   RP11_FP RP11_be AC_be   AS_be   Info
AE02    0.060233    0   0.682884    0.817115    0.591182    0.129252    SAP
AE03    0   0   0   0.889181    0.670113    0.766243    SAP
AE04    0   0   0.033256    0.726193    0.171861    0.103839    others
AE05    0   0   0.034988    0.451329    0.431836    0.219843    others

What I am aiming is to plot each column starting from AS_FP til RP11_beta as lmplot, each x axis is column ending with FP and y axis is its corresponding column ending with be. And I wanted to save it as separate files so I strated iterating through the columns by skipping first column ID, like this,

for ind, column in enumerate(df.columns):
    if column.split('_')[0] == column.split('_')[0]:

But I got lost how to continue, I need to plot

sns.lmplot(x, y, data=df, hue='Info',palette=colors, fit_reg=False,
           size=10,scatter_kws={"s": 700},markers=["o", "v"])

and save each image as seperate file

Alexey Trofimov · Accepted Answer · 2017-09-25 12:55:36Z

1

Straightforward solution:

1) Toy data:

import pandas as pd
from collections import OrderedDict
import matplotlib.pyplot as plt
import seaborn as sns

dct = OrderedDict()
dct["ID"] = ["AE02", "AE03", "AE04", "AE05"]
dct["AS_FP"] = [0.060233, 0, 0, 0]
dct["AC_FP"] = [0, 0,0, 0]
dct["RP11_FP"] = [0.682884, 0, 0.033256, 0.034988]
dct["AS_be"] = [0.129252, 0.766243, 0.103839, 0.219843]
dct["AC_be"] = [0.591182, 0.670113, 0.171861, 0.431836]
dct["RP11_be"] = [0.817115, 0.889181, 0.726193, 0.451329]
dct["Info"] = ["SAP", "SAP", "others", "others"]

df = pd.DataFrame(dct)

2) Iterating through pairs, saving each figure with unique filename:

graph_cols = [col for col in df.columns if ("_FP" in col) or ("_be" in col)]

fps = sorted([col for col in graph_cols if "_FP" in col])
bes = sorted([col for col in graph_cols if "_be" in col])

for x, y in zip(fps, bes):
    snsplot = sns.lmplot(x, y, data=df, fit_reg=False, hue='Info',
           size=10, scatter_kws={"s": 700})
    snsplot.savefig(x.split("_")[0] + ".png")

You can add needed params in lmlplot as you need.

edited Sep 25, 2017 at 12:55

answered Sep 25, 2017 at 12:34

Alexey Trofimov

5,0672 gold badges24 silver badges30 bronze badges

Sign up to request clarification or add additional context in comments.

9 Comments

ARJ Over a year ago

Thank you for the solution, actually the dimension of my data frame is (82, 28), and so when i try the solution I get the following error message, (operands could not be broadcast together with shapes (2,) (82,) )

Alexey Trofimov Over a year ago

So please tell us more about your data if your toy example doesnt represent it. I guess you have some other columns except for pairs and ID. Ive added column check in the answer, but the error youve mentioned is numpy exception, so adittional info is needed.

ARJ Over a year ago

Yes thats true. I have an additional column Info which I need to use for hue in lmplot. Additionally the columns are not dictionary ordered as the toy dataframe you showed. I have edited my primary question. Please have a look

Alexey Trofimov Over a year ago

Ive updated the answer with hue. Ordered dict doesnt make sense here, its just for convinience(you can make it unordered with dct={} ). We sort lthe corresponding lists to get right pairs.

ARJ Over a year ago

Thank you, its not possible as graph_col is a list I am not able to sort it . AttributeError: 'list' object has no attribute 'columns'

|

Collectives™ on Stack Overflow

Iterating over columns in data frame by skipping first column and drawing multiple plots

1 Answer 1

9 Comments

Your Answer

Hot Network Questions

Collectives™ on Stack Overflow

1 Answer 1

9 Comments

Your Answer

Sign up or log in

Post as a guest

Related