Chapter 2 Use of fbprohet

The following code is analyzed under jupyter, the process

import pandas as pd
from prophet import Prophet

Read data using pandas
df = pd.read_csv('./examples/example_wp_log_peyton_manning.csv')

Create a Prophet() object
m = Prophet()

Initialize the data, the fit method builds a model through the given data

m.fit(df)

Specify the prediction range
future = m.make_future_dataframe(periods=365)

Execute forecast
forecast = m.predict(future)

Display the forecast results for the original data
fig1 = m.plot(forecast)

Because the x coordinates, that is, the ds column, are in datetime format, this function is divided according to time

Trend, holiday, week, season time periods for analysis

fig2 = m.plot_components(forecast)

display data

View all column data
in jupyter
forecast
forecast = Prophet object.fit(df).predict(future)
display
fig = Prophet object.plot_components(forecast)

View data changes throughout the year
from prophet.plot import plot_yearly
m = Prophet().fit(df)
a = plot_yearly(m)

def plot_yearly(m, ax=None, uncertainty=True,
yearly_start=0, figsize=(10, 6), name=‘yearly’):

Plot the annual components of the forecast.

 参数
 ----------
 m：先知模型。
 ax：要绘制的可选 matplotlib 轴。 一个将被创建，如果
     这不提供。
 不确定性：可选布尔值来绘制不确定性区间，这将
     仅在 m.uncertainty_samples > 0 时执行。
 yearly_start: 可选 int 指定每年的开始日期
     季节性图。 0（默认）从 1 月 1 日开始这一年。1 次轮班
     从 1 天到 1 月 2 日，依此类推。
 figsize: 可选的元组宽度，以英寸为单位的高度。
 name：季节性组件的名称（如果之前从默认的“每年”更改）。
 -------

Simply look at the data, showing the uncertainty interval
df = pd.read_csv('.../examples/example_wp_log_R_outliers1.csv')
m = Prophet()
m.fit(df)
future = m.make_future_dataframe(periods=1096)
forecast = m.predict (future)
fig = m.plot(forecast)