CODEWITHSUNDEEP
pythonpandas
Steve
Steve

Reputation: 331

Grouping a Pandas Dataframe by Date with Python and Plotting

I have a comma separated file that I am reading with Pandas via Python. Here is a subset: 

Location   Time                 Value1 Value2 
CCNY       2013-01-01 00:00:00  59     12.71    
CCNY       2013-01-01 01:00:00  96     10.6 
CCNY       2013-01-01 02:00:00  105    11.94    
CCNY       2013-01-01 03:00:00  81     11.73    
CCNY       2013-01-01 04:00:00  60     13.05    
CCNY       2013-01-01 05:00:00  51     13.25
... 
CCNY       2013-31-01 06:00:00  28     13.03

I need to plot value1 (x-axis) vs value2 (y-axis), but I need to do it for each day. So, for this portion of the file which contains values for the entire month of January, there will be 31 plots.

How should I go about this?

(The ultimate goal is to get best fit lines in each plot and r squared values.)

Thanks.

Upvotes: 2

Views: 76

Answers (2)

EdChum
EdChum

Reputation: 393893

Your csv looks like a fixed width file so I would use read_fwf, you then need to rename the time column as it treats the time portion as an unnamed column, you can then use @chrisB's answer to achieve what you want:

In [35]:

t="""Location   Time                 Value1 Value2 
CCNY       2013-01-01 00:00:00  59     12.71    
CCNY       2013-01-01 01:00:00  96     10.6 
CCNY       2013-01-01 02:00:00  105    11.94    
CCNY       2013-01-02 03:00:00  81     11.73    
CCNY       2013-01-02 04:00:00  60     13.05    
CCNY       2013-01-02 05:00:00  51     13.25"""
​
df = pd.read_fwf(io.StringIO(t), parse_dates=[[1,2]])
df.rename(columns={'Time_Unnamed: 2':'Time'},inplace=True)
df
Out[35]:
                 Time Location  Value1  Value2
0 2013-01-01 00:00:00     CCNY      59   12.71
1 2013-01-01 01:00:00     CCNY      96   10.60
2 2013-01-01 02:00:00     CCNY     105   11.94
3 2013-01-02 03:00:00     CCNY      81   11.73
4 2013-01-02 04:00:00     CCNY      60   13.05
5 2013-01-02 05:00:00     CCNY      51   13.25
In [36]:

df.groupby(df['Time'].dt.date).plot(x='Value1', y='Value2')
Out[36]:
2013-01-01    Axes(0.125,0.125;0.775x0.775)
2013-01-02    Axes(0.125,0.125;0.775x0.775)
dtype: object

Results in the plots:

enter image description here

and

enter image description here

Upvotes: 1

chrisb
chrisb

Reputation: 52236

See below - this groups the data by day and produce a plot for each.

df.groupby(df['Time'].dt.day).plot(x='Value1', y='Value2')

Upvotes: 3

Related Questions