Predicting The Future with Python

PYMNTOs, June 8, 2017

In [1]:
import pandas as pd
import numpy as np
from fbprophet import Prophet
In [2]:
%matplotlib inline
In [3]:
# Read in the source data - downloaded from google analytics
df = pd.read_excel('https://github.com/chris1610/pbpython/blob/master/data/All-Web-Site-Data-Audience-Overview.xlsx?raw=True')
In [4]:
df.head()
Out[4]:
Day Index Sessions
0 2014-09-25 1
1 2014-09-26 4
2 2014-09-27 8
3 2014-09-28 42
4 2014-09-29 233
In [5]:
# Convert to log format
df['Sessions'] = np.log(df['Sessions'])
In [6]:
# Need to name the columns like this in order for prophet to work
df.columns = ["ds", "y"]
In [7]:
# Create the model
m1 = Prophet()
m1.fit(df)
Out[7]:
<fbprophet.forecaster.Prophet at 0x7f254b0a7d68>
In [8]:
# Predict out a year
future1 = m1.make_future_dataframe(periods=365)
forecast1 = m1.predict(future1)
In [9]:
m1.plot(forecast1);
In [10]:
m1.plot_components(forecast1);
In [11]:
articles = pd.DataFrame({
  'holiday': 'publish',
  'ds': pd.to_datetime(['2014-09-27', '2014-10-05', '2014-10-14', '2014-10-26', '2014-11-9',
                        '2014-11-18', '2014-11-30', '2014-12-17', '2014-12-29', '2015-01-06',
                        '2015-01-20', '2015-02-02', '2015-02-16', '2015-03-23', '2015-04-08',
                        '2015-05-04', '2015-05-17', '2015-06-09', '2015-07-02', '2015-07-13',
                        '2015-08-17', '2015-09-14', '2015-10-26', '2015-12-07', '2015-12-30',
                        '2016-01-26', '2016-04-06', '2016-05-16', '2016-06-15', '2016-08-23',
                        '2016-08-29', '2016-09-06', '2016-11-21', '2016-12-19', '2017-01-17',
                        '2017-02-06', '2017-02-21', '2017-03-06']),
  'lower_window': 0,
  'upper_window': 5,
})
articles.head()
Out[11]:
ds holiday lower_window upper_window
0 2014-09-27 publish 0 5
1 2014-10-05 publish 0 5
2 2014-10-14 publish 0 5
3 2014-10-26 publish 0 5
4 2014-11-09 publish 0 5
In [12]:
m2 = Prophet(holidays=articles).fit(df)
future2 = m2.make_future_dataframe(periods=365)
forecast2 = m2.predict(future2)
m2.plot(forecast2);
In [13]:
m2.plot_components(forecast2);
In [14]:
forecast2.head()
Out[14]:
ds t trend seasonal_lower seasonal_upper trend_lower trend_upper yhat_lower yhat_upper publish publish_lower publish_upper weekly weekly_lower weekly_upper yearly yearly_lower yearly_upper seasonal yhat
0 2014-09-25 0.000000 3.080198 0.072991 0.072991 3.080198 3.080198 2.608365 3.742410 0.000000 0.000000 0.000000 0.162671 0.162671 0.162671 -0.089680 -0.089680 -0.089680 0.072991 3.153189
1 2014-09-26 0.001124 3.110246 -0.114283 -0.114283 3.110246 3.110246 2.470895 3.550829 0.000000 0.000000 0.000000 0.002165 0.002165 0.002165 -0.116448 -0.116448 -0.116448 -0.114283 2.995963
2 2014-09-27 0.002247 3.140294 -0.103136 -0.103136 3.140294 3.140294 2.512293 3.563757 0.532607 0.532607 0.532607 -0.493243 -0.493243 -0.493243 -0.142501 -0.142501 -0.142501 -0.103136 3.037158
3 2014-09-28 0.003371 3.170343 0.404199 0.404199 3.170343 3.170343 3.034737 4.100938 0.883361 0.883361 0.883361 -0.312015 -0.312015 -0.312015 -0.167148 -0.167148 -0.167148 0.404199 3.574542
4 2014-09-29 0.004494 3.200391 0.528059 0.528059 3.200391 3.200391 3.185862 4.290530 0.427628 0.427628 0.427628 0.290137 0.290137 0.290137 -0.189706 -0.189706 -0.189706 0.528059 3.728450
In [15]:
forecast2["Sessions"] = np.exp(forecast2.yhat).round()
In [19]:
forecast2[["ds", "Sessions"]].tail()
Out[19]:
ds Sessions
1251 2018-02-27 1998.0
1252 2018-02-28 1911.0
1253 2018-03-01 1905.0
1254 2018-03-02 1607.0
1255 2018-03-03 970.0
In [ ]: