vastorbit.VastFrame.pacf¶

VastFrame.pacf(column: str, ts: str, by: Annotated[str | list[str], 'STRING representing one column or a list of columns'] | None = None, p: int | list = 5, unit: str = 'rows', method: Literal['pearson', 'kendall', 'spearman', 'spearmand', 'biserial', 'cramer'] = 'pearson', confidence: bool = True, alpha: float = 0.95, show: bool = True, kind: Literal['line', 'bar'] = 'bar', chart: PlottingBase | TableSample | Axes | mFigure | Figure | None = None, **style_kwargs)¶

Computes the partial autocorrelations of the specified VastColumn. Partial autocorrelations are a fundamental concept in time series analysis and provide essential information about the dependencies between data points at different time lags. Understanding these partial autocorrelations can aid in modeling and predicting future values, making it a valuable tool for time series analysis and forecasting.

Parameters:

column (str) – Input VastColumn used to compute the Auto Correlation Plot.
ts (str) – TS (Time Series) VastColumn used to order the data. It can be of type date or a numerical VastColumn.
by (SQLColumns, optional) – VastColumns used in the partition.
p (int | list, optional) – Int equal to the maximum number of lag to consider during the computation or List of the different lags to include during the computation. p must be positive or a list of positive integers.
unit (str, optional) –
Unit used to compute the lags.
- rows:
  Natural lags.
- else:
  Any time unit. For example, you can write ‘hour’ to compute the hours lags or ‘day’ to compute the days lags.
method (str, optional) –
Method used to compute the correlation.
- pearson:
  Pearson’s correlation coefficient (linear).
- spearman:
  Spearman’s correlation coefficient (monotonic - rank based).
- spearmanD:
  Spearman’s correlation coefficient using the DENSE RANK function instead of the RANK function.
- kendall:
  Kendall’s correlation coefficient (similar trends). The method computes the Tau-B coefficient.
  
  Warning
  
  This method uses a CROSS JOIN during computation and is therefore computationally expensive at O(n * n), where n is the total count of the VastFrame.
- cramer:
  Cramer’s V (correlation between categories).
- biserial:
  Biserial Point (correlation between binaries and a numericals).
confidence (bool, optional) – If set to True, the confidence band width is drawn.
alpha (float, optional) – Significance Level. Probability to accept H0. Only used to compute the confidence band width.
show (bool, optional) – If set to True, the Plotting object is returned.
kind (str, optional) –
PACF Type.
- bar:
  Classical Partial Autocorrelation Plot using bars.
- line:
  Draws the PACF using a Line Plot.
chart (PlottingObject, optional) – The chart object used to plot.
**style_kwargs – Any optional parameter to pass to the plotting functions.

Returns:

Plotting Object.

Return type:

obj

Examples

Import the amazon dataset from vastorbit.

from vastorbit.datasets import load_amazon

data = load_amazon()

	📅 date Date	Abc state Varchar(32)	123 number Integer
1	2016-01-01	PARAÍBA	18
2	2016-02-01	PARAÍBA	4
3	2016-03-01	PARAÍBA	1
4	2016-04-01	PARAÍBA	1
5	2016-05-01	PARAÍBA	1
6	2016-06-01	PARAÍBA	4
7	2016-07-01	PARAÍBA	22
8	2016-08-01	PARAÍBA	50
9	2016-09-01	PARAÍBA	131
10	2016-10-01	PARAÍBA	304
11	2016-11-01	PARAÍBA	132
12	2016-12-01	PARAÍBA	8
13	2016-01-01	PARÁ	1322
14	2016-02-01	PARÁ	430
15	2016-03-01	PARÁ	81
16	2016-04-01	PARÁ	66
17	2016-05-01	PARÁ	154
18	2016-06-01	PARÁ	502
19	2016-07-01	PARÁ	1579
20	2016-08-01	PARÁ	4863
21	2016-09-01	PARÁ	3953
22	2016-10-01	PARÁ	5281
23	2016-11-01	PARÁ	7879
24	2016-12-01	PARÁ	3300
25	2016-01-01	PERNAMBUCO	24
26	2016-02-01	PERNAMBUCO	19
27	2016-03-01	PERNAMBUCO	4
28	2016-04-01	PERNAMBUCO	8
29	2016-05-01	PERNAMBUCO	6
30	2016-06-01	PERNAMBUCO	4
31	2016-07-01	PERNAMBUCO	17
32	2016-08-01	PERNAMBUCO	42
33	2016-09-01	PERNAMBUCO	171
34	2016-10-01	PERNAMBUCO	319
35	2016-11-01	PERNAMBUCO	191
36	2016-12-01	PERNAMBUCO	161
37	2016-01-01	PIAUÍ	94
38	2016-02-01	PIAUÍ	97
39	2016-03-01	PIAUÍ	53
40	2016-04-01	PIAUÍ	35
41	2016-05-01	PIAUÍ	60
42	2016-06-01	PIAUÍ	153
43	2016-07-01	PIAUÍ	754
44	2016-08-01	PIAUÍ	1647
45	2016-09-01	PIAUÍ	1394
46	2016-10-01	PIAUÍ	2598
47	2016-11-01	PIAUÍ	1126
48	2016-12-01	PIAUÍ	374
49	2016-01-01	RIO DE JANEIRO	9
50	2016-02-01	RIO DE JANEIRO	16
51	2016-03-01	RIO DE JANEIRO	16
52	2016-04-01	RIO DE JANEIRO	45
53	2016-05-01	RIO DE JANEIRO	30
54	2016-06-01	RIO DE JANEIRO	37
55	2016-07-01	RIO DE JANEIRO	131
56	2016-08-01	RIO DE JANEIRO	241
57	2016-09-01	RIO DE JANEIRO	195
58	2016-10-01	RIO DE JANEIRO	30
59	2016-11-01	RIO DE JANEIRO	19
60	2016-12-01	RIO DE JANEIRO	5
61	2016-01-01	RIO GRANDE DO NORTE	15
62	2016-02-01	RIO GRANDE DO NORTE	2
63	2016-03-01	RIO GRANDE DO NORTE	1
64	2016-04-01	RIO GRANDE DO NORTE	0
65	2016-05-01	RIO GRANDE DO NORTE	1
66	2016-06-01	RIO GRANDE DO NORTE	4
67	2016-07-01	RIO GRANDE DO NORTE	13
68	2016-08-01	RIO GRANDE DO NORTE	24
69	2016-09-01	RIO GRANDE DO NORTE	44
70	2016-10-01	RIO GRANDE DO NORTE	129
71	2016-11-01	RIO GRANDE DO NORTE	93
72	2016-12-01	RIO GRANDE DO NORTE	75
73	2016-01-01	RIO GRANDE DO SUL	68
74	2016-02-01	RIO GRANDE DO SUL	55
75	2016-03-01	RIO GRANDE DO SUL	30
76	2016-04-01	RIO GRANDE DO SUL	32
77	2016-05-01	RIO GRANDE DO SUL	37
78	2016-06-01	RIO GRANDE DO SUL	261
79	2016-07-01	RIO GRANDE DO SUL	865
80	2016-08-01	RIO GRANDE DO SUL	1111
81	2016-09-01	RIO GRANDE DO SUL	628
82	2016-10-01	RIO GRANDE DO SUL	93
83	2016-11-01	RIO GRANDE DO SUL	105
84	2016-12-01	RIO GRANDE DO SUL	79
85	2016-01-01	RONDÔNIA	93
86	2016-02-01	RONDÔNIA	88
87	2016-03-01	RONDÔNIA	25
88	2016-04-01	RONDÔNIA	59
89	2016-05-01	RONDÔNIA	44
90	2016-06-01	RONDÔNIA	170
91	2016-07-01	RONDÔNIA	969
92	2016-08-01	RONDÔNIA	3675
93	2016-09-01	RONDÔNIA	4208
94	2016-10-01	RONDÔNIA	1844
95	2016-11-01	RONDÔNIA	401
96	2016-12-01	RONDÔNIA	148
97	2016-01-01	RORAIMA	1754
98	2016-02-01	RORAIMA	171
99	2016-03-01	RORAIMA	1081
100	2016-04-01	RORAIMA	126

Rows: 1-100 | Columns: 3

Draw the PACF Plot.

data.pacf(
    column = "number",
    ts = "date",
    by = "state",
    method = "pearson",
    p = 24,
)

For more examples, please look at the Auto-Correlation Plot page of the Chart Gallery. Those ones are related to ACF plots, but the customization stays the same for the PACF plot.