Script to extract ferret timeseries data into two columns of numbers
In [1]:
import pandas as pd
# For bokeh plots
import bokeh #0.12.3
from bokeh.plotting import figure
from datetime import datetime as dt
from bokeh.models import DatetimeTickFormatter
from math import pi
In [2]:
# Read all csv files and create a dataframe from selected year and selected fields
# df = pd.DataFrame()
# for key in dict_files.keys():
# df_full = pd.read_csv('files_csv/' + dict_files[key], delimiter=';', index_col=0)
# df_select = df_full.loc[year][list_fields]
# df_select['Source'] = key
# df = df.append(df_select)
# df.set_index('Source', inplace=True)
# df
In [3]:
# Read csv file output by ferret
df_orig = pd.read_csv('~/timeseries.csv', delimiter=':', index_col=0)
#df.set_index('Source', inplace=True)
df_orig.head(10)
Out[3]:
In [4]:
# Read csv file output by ferret, skipping header
df_full = pd.read_csv('~/timeseries.csv', delimiter=':', skiprows=5)
#df.set_index('Source', inplace=True)
df_full.head(5)
Out[4]:
In [5]:
#Extract date string from col 0
s = df_full.ix[:, 0]
s.head(5)
Out[5]:
In [6]:
# x_tmp = s.str.split(' /').str.get(0)
# x_tmp.head(5)
In [18]:
x = s.str.split(' /').str.get(0).str.split(' ').str.get(1)
x.head(12)
Out[18]:
In [8]:
x[0]
Out[8]:
In [25]:
# Determine if x-col is a list of dates
months=['JAN', 'FEB', 'MAR', 'APR', 'MAY', 'JUN', 'JUL', 'AUG', 'SEP', 'OCT', 'NOV', 'DEC']
dateFlag=0
for idx in range(12):
if x[0].find(months[idx]) != -1:
print("found month")
dateFlag=1
print dateFlag
In [27]:
if dateFlag==1:
x = pd.to_datetime(x,infer_datetime_format=True)
x.head(5)
Out[27]:
In [29]:
# Put x and y values in dataframe
dfer = pd.DataFrame()
dfer['xval'] = x
dfer['yval'] = df_full.ix[:, 1]
dfer.head(5)
Out[29]:
In [39]:
# dfer['date'] = pd.to_datetime(dfer['date'], format='%d%b%Y:%H:%M:%S.%f')
In [31]:
p = figure(title='A Bokeh plot',
plot_width=700,plot_height=400)
In [32]:
p.line(dfer['xval'], dfer['yval'])
Out[32]:
In [33]:
if dateFlag==1:
p.xaxis.formatter=DatetimeTickFormatter(formats=dict(
hours=["%d %B %Y"],
days=["%d %B %Y"],
months=["%d %B %Y"],
years=["%d %B %Y"],
))
p.xaxis.major_label_orientation = pi/4
p.yaxis.axis_label = "size"
In [34]:
# For plotting in notebook
from bokeh.io import output_file
from bokeh.charts import show
from bokeh.models import DatetimeTickFormatter
output_file('myplot2.html')
show(p)
In [ ]:
In [29]:
# http://stackoverflow.com/questions/33869292/how-can-i-set-the-x-axis-as-datetimes-on-a-bokeh-plot
import pandas as pd
from math import pi
from datetime import datetime as dt
from bokeh.io import output_file
from bokeh.charts import show
from bokeh.models import DatetimeTickFormatter
from bokeh.plotting import figure
df = pd.DataFrame(data=[1,2,3],
index=[dt(2015, 1, 1), dt(2015, 1, 2), dt(2015, 1, 3)],
columns=['foo'])
p_test = figure(plot_width=400, plot_height=400)
p_test.line(df.index, df['foo'])
p_test.xaxis.formatter=DatetimeTickFormatter(formats=dict(
hours=["%d %B %Y"],
days=["%d %B %Y"],
months=["%d %B %Y"],
years=["%d %B %Y"],
))
p_test.xaxis.major_label_orientation = pi/4
output_file('myplot.html')
show(p_test)
In [95]:
df.index
Out[95]:
In [118]:
df['foo']
Out[118]:
In [30]:
p_test.line(df.index, df['foo'])
Out[30]:
In [ ]: