0

So, I've made a graph that plots multiple rows from a .csv file onto one graph. I want to make another graph that only includes specific rows to better analyze them.

My data is (only a portion of the total):

Number  Base-BW  1-BW    2-BW   3-BW    4-End-BW
182-14  39.7    41.1    40.3    39.5    38.8
95-14   43.3    41      41.9    42.4    41.6
15-14   59.4    59.4    59.1    59.1    56.4
124-14  76.4    77.4    74.8    74.5    68.1
183-14  35      35.5    36.8    37.3    35.4

My current script is:

# import all needed modules
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
%matplotlib inline
# Set some Pandas options
import matplotlib.colors as cl
cl.Colormap('hsv')
plt.plot([0,1,2,3,4], (df['Base-BW'],df['1-BW'],df['2-BW'],df['3-BW'],df['4-End-BW']))
plt.axis([0, 4, 0, 100])
plt.xlabel('Day')
plt.ylabel('Weight (g)')
plt.title('Body Weight')
plt.legend(df['Number'], bbox_to_anchor=(1.1, 1.01))
plt.show()

Which yields:

enter image description here

So is there a way to just show, say, row1 (182-14), 3 (15-14), and 5 (183-14)? Or maybe just to graph those specific rows? I've checked other questions, and none pertain to using a .csv file like I am.

Michael S.
  • 3,050
  • 4
  • 19
  • 34
  • might help, [this answer](http://stackoverflow.com/questions/13428318/reading-rows-from-a-csv-file-in-python) – Srivatsan Feb 27 '15 at 21:13
  • 1
    You're not showing entirety of your code, what is `SelectBreed` how exactly do you read in your data? (I don't see a function or a csv reader that does that for you). Generally there is no universal function that does that. Your best bet is to read in csv file and store each line in a list, i.e. `line`, as a list and then plot it like `plt.plot(x_values, line[i]` where x values seem to be days (although your code plots a `[1,2,3,4]` as days. I might be missing something, but it doesn't look like that code should produce your result. Then again it depends a lot on what `SelectBreed` is. – ljetibo Feb 27 '15 at 21:18
  • I updated it to include the entire code, I didn't know if that was necessary or not. Still kinda new to python. – Michael S. Feb 27 '15 at 21:22
  • 1
    If you want us to test *your* data, making it as easy as possible to reconstitute it on our machines will help -- usually you can write out a section of the data structure as a string that can be re-read without, e.g., saving a file to disk. Which is a pain. – cphlewis Feb 27 '15 at 21:24

1 Answers1

1

The Pandas library rocks for reading .csv files and then analyzing them. Using a sample dataset from R, 

import pandas as pd
cw = pd.read_csv('ChickWeight.csv')
cw.columns

Out[166]: Index([u'weight', u'Time', u'Chick', u'Diet'], dtype='object')

cw[cw.Chick==1].plot('Time','weight')

enter image description here

Picking apart that last line: cw[cw.Chick==1] is all rows of cw for which Chick==1; pandas dataframes have a plot method (which is matplotlib, I think, so will be familiar).

You can also do complicated selections from the dataframe, there are some good related pandas questions.

edited to add: use a list of values to select rows from a pandas dataframe is probably the most literal approach to your question; Selecting rows from a Pandas dataframe with a compound (hierarchical) index is pretty cool for doing selections based on more than one column (comes up all the time in multi-factor experiments).

Community
  • 1
  • 1
cphlewis
  • 15,759
  • 4
  • 46
  • 55