Other Utilities

Generating Full Span of multiple time-series

import pandas as pd
import numpy as np
from orbit.utils.general import expand_grid, regenerate_base_df

Define the series keys and datetime array.

dt = pd.date_range('2020-01-31', '2022-12-31', freq='M')
keys = ['x' + str(x) for x in range(10)]
print(keys)
print(dt)

['x0', 'x1', 'x2', 'x3', 'x4', 'x5', 'x6', 'x7', 'x8', 'x9']
DatetimeIndex(['2020-01-31', '2020-02-29', '2020-03-31', '2020-04-30',
               '2020-05-31', '2020-06-30', '2020-07-31', '2020-08-31',
               '2020-09-30', '2020-10-31', '2020-11-30', '2020-12-31',
               '2021-01-31', '2021-02-28', '2021-03-31', '2021-04-30',
               '2021-05-31', '2021-06-30', '2021-07-31', '2021-08-31',
               '2021-09-30', '2021-10-31', '2021-11-30', '2021-12-31',
               '2022-01-31', '2022-02-28', '2022-03-31', '2022-04-30',
               '2022-05-31', '2022-06-30', '2022-07-31', '2022-08-31',
               '2022-09-30', '2022-10-31', '2022-11-30', '2022-12-31'],
              dtype='datetime64[ns]', freq='M')

Users can use expand_grid to generate dataframe with observations in key and dt levels.

df_base = expand_grid({
    'key': keys,
    'dt': dt,
})
x = np.random.normal(0, 1, 10 * 36)
df_base['x'] = x
print(df_base.shape)
df_base.head(5)

(360, 3)

	key	dt	x
0	x0	2020-01-31	-0.588512
1	x0	2020-02-29	1.547821
2	x0	2020-03-31	1.114709
3	x0	2020-04-30	1.410516
4	x0	2020-05-31	-0.699229

Regenerate Multiple Timeseries with Missing rows

Create missing rows.

np.random.seed(2022)
drop_idx = np.random.choice(df_base.index, 5, replace=False)
df_missing = df_base.drop(drop_idx).reset_index(drop=True)
print(df_missing.shape)
df_missing.head(5)

(355, 3)

	key	dt	x
0	x0	2020-01-31	-0.588512
1	x0	2020-02-29	1.547821
2	x0	2020-03-31	1.114709
3	x0	2020-04-30	1.410516
4	x0	2020-05-31	-0.699229

Use regenerate_base_df to regenerate the base dataframe.

time_col = "dt"
key_col = "key"
new_df_base = regenerate_base_df(df_missing, time_col, key_col, val_cols=['x'])

By default, the missing entries regenerated come with a null value.

new_df_base.iloc[drop_idx]

	dt	key	x
286	2022-11-30	x7	NaN
274	2021-11-30	x7	NaN
75	2020-04-30	x2	NaN
135	2022-04-30	x3	NaN
43	2020-08-31	x1	NaN

Users can also use fill_na option to fill the missing values.

new_df_base = regenerate_base_df(df_missing, time_col, key_col, val_cols=['x'], fill_na=0)

new_df_base.iloc[drop_idx]

	dt	key	x
286	2022-11-30	x7	0.0
274	2021-11-30	x7	0.0
75	2020-04-30	x2	0.0
135	2022-04-30	x3	0.0
43	2020-08-31	x1	0.0