VerticaPy

Python API for Vertica Data Science at Scale

Correlation and Dependency¶

Finding links between variables is a very important task. The main purpose of data science is to find relationships between variables, and to understand how these relationships can help us make better decisions.

Machine learning models are also sensitive to the number of variables and how they relate and affect each other, so finding correlations and dependencies can help us make better use of our machine learning algorithms.

Let's use the Telco Churn dataset to understand how we can find links between different variables in VerticaPy.

In [1]:

import verticapy as vp
vdf = vp.read_csv("data/churn.csv")
display(vdf)

	Abc customerID Varchar(20)	Abc gender Varchar(20)	123 SeniorCitizen Int	010 Partner Boolean	010 Dependents Boolean	123 tenure Int	010 PhoneService Boolean	Abc MultipleLines Varchar(100)	Abc InternetService Varchar(22)	Abc OnlineSecurity Varchar(38)	Abc OnlineBackup Varchar(38)	Abc DeviceProtection Varchar(38)	Abc TechSupport Varchar(38)	Abc StreamingTV Varchar(38)	Abc StreamingMovies Varchar(38)	Abc Contract Varchar(28)	010 PaperlessBilling Boolean	Abc PaymentMethod Varchar(50)	123 MonthlyCharges Numeric(8,3)	123 TotalCharges Numeric(9,3)	010 Churn Boolean
1	0002-ORFBO	Female	0	✅	✅	9	✅	No	DSL	No	Yes	No	Yes	Yes	No	One year	✅	Mailed check	65.6	593.3	❌
2	0003-MKNFE	Male	0	❌	❌	9	✅	Yes	DSL	No	No	No	No	No	Yes	Month-to-month	❌	Mailed check	59.9	542.4	❌
3	0004-TLHLJ	Male	0	❌	❌	4	✅	No	Fiber optic	No	No	Yes	No	No	No	Month-to-month	✅	Electronic check	73.9	280.85	✅
4	0011-IGKFF	Male	1	✅	❌	13	✅	No	Fiber optic	No	Yes	Yes	No	Yes	Yes	Month-to-month	✅	Electronic check	98.0	1237.85	✅
5	0013-EXCHZ	Female	1	✅	❌	3	✅	No	Fiber optic	No	No	No	Yes	Yes	No	Month-to-month	✅	Mailed check	83.9	267.4	✅
6	0013-MHZWF	Female	0	❌	✅	9	✅	No	DSL	No	No	No	Yes	Yes	Yes	Month-to-month	✅	Credit card (automatic)	69.4	571.45	❌
7	0013-SMEOE	Female	1	✅	❌	71	✅	No	Fiber optic	Yes	Yes	Yes	Yes	Yes	Yes	Two year	✅	Bank transfer (automatic)	109.7	7904.25	❌
8	0014-BMAQU	Male	0	✅	❌	63	✅	Yes	Fiber optic	Yes	No	No	Yes	No	No	Two year	✅	Credit card (automatic)	84.65	5377.8	❌
9	0015-UOCOJ	Female	1	❌	❌	7	✅	No	DSL	Yes	No	No	No	No	No	Month-to-month	✅	Electronic check	48.2	340.35	❌
10	0016-QLJIS	Female	0	✅	✅	65	✅	Yes	DSL	Yes	Yes	Yes	Yes	Yes	Yes	Two year	✅	Mailed check	90.45	5957.9	❌
11	0017-DINOC	Male	0	❌	❌	54	❌	No phone service	DSL	Yes	No	No	Yes	Yes	No	Two year	❌	Credit card (automatic)	45.2	2460.55	❌
12	0017-IUDMW	Female	0	✅	✅	72	✅	Yes	Fiber optic	Yes	Yes	Yes	Yes	Yes	Yes	Two year	✅	Credit card (automatic)	116.8	8456.75	❌
13	0018-NYROU	Female	0	✅	❌	5	✅	No	Fiber optic	No	No	No	No	No	No	Month-to-month	✅	Electronic check	68.95	351.5	❌
14	0019-EFAEP	Female	0	❌	❌	72	✅	Yes	Fiber optic	Yes	Yes	Yes	No	Yes	No	Two year	✅	Bank transfer (automatic)	101.3	7261.25	❌
15	0019-GFNTW	Female	0	❌	❌	56	❌	No phone service	DSL	Yes	Yes	Yes	Yes	No	No	Two year	❌	Bank transfer (automatic)	45.05	2560.1	❌
16	0020-INWCK	Female	0	✅	✅	71	✅	Yes	Fiber optic	No	Yes	Yes	No	No	Yes	Two year	✅	Credit card (automatic)	95.75	6849.4	❌
17	0020-JDNXP	Female	0	✅	✅	34	❌	No phone service	DSL	Yes	No	Yes	Yes	Yes	Yes	One year	❌	Mailed check	61.25	1993.2	❌
18	0021-IKXGC	Female	1	❌	❌	1	✅	Yes	Fiber optic	No	No	No	No	No	No	Month-to-month	✅	Electronic check	72.1	72.1	❌
19	0022-TCJCI	Male	1	❌	❌	45	✅	No	DSL	Yes	No	Yes	No	No	Yes	One year	❌	Credit card (automatic)	62.7	2791.5	✅
20	0023-HGHWL	Male	1	❌	❌	1	❌	No phone service	DSL	No	No	No	No	No	No	Month-to-month	✅	Electronic check	25.1	25.1	✅
21	0023-UYUPN	Female	1	✅	❌	50	✅	Yes	No	No internet service	No internet service	No internet service	No internet service	No internet service	No internet service	One year	❌	Electronic check	25.2	1306.3	❌
22	0023-XUOPT	Female	0	✅	❌	13	✅	Yes	Fiber optic	No	Yes	Yes	No	Yes	No	Month-to-month	❌	Electronic check	94.1	1215.6	✅
23	0027-KWYKW	Female	0	✅	✅	23	✅	Yes	Fiber optic	No	No	No	No	Yes	No	Month-to-month	✅	Electronic check	83.75	1849.95	❌
24	0030-FNXPP	Female	0	❌	❌	3	✅	No	No	No internet service	No internet service	No internet service	No internet service	No internet service	No internet service	Month-to-month	❌	Mailed check	19.85	57.2	❌
25	0031-PVLZI	Female	0	✅	✅	4	✅	No	No	No internet service	No internet service	No internet service	No internet service	No internet service	No internet service	Month-to-month	❌	Mailed check	20.35	76.35	✅
26	0032-PGELS	Female	0	✅	✅	1	❌	No phone service	DSL	Yes	No	No	No	No	No	Month-to-month	❌	Bank transfer (automatic)	30.5	30.5	✅
27	0036-IHMOT	Female	0	✅	✅	55	✅	No	Fiber optic	No	Yes	Yes	Yes	Yes	Yes	One year	✅	Bank transfer (automatic)	103.7	5656.75	❌
28	0040-HALCW	Male	0	✅	✅	54	✅	No	No	No internet service	No internet service	No internet service	No internet service	No internet service	No internet service	Two year	❌	Credit card (automatic)	20.4	1090.6	❌
29	0042-JVWOJ	Male	0	❌	❌	26	✅	No	No	No internet service	No internet service	No internet service	No internet service	No internet service	No internet service	One year	✅	Bank transfer (automatic)	19.6	471.85	❌
30	0042-RLHYP	Female	0	✅	✅	69	✅	No	No	No internet service	No internet service	No internet service	No internet service	No internet service	No internet service	Two year	❌	Bank transfer (automatic)	19.7	1396.9	❌
31	0048-LUMLS	Male	0	✅	✅	37	✅	No	Fiber optic	No	No	No	No	Yes	Yes	One year	❌	Credit card (automatic)	91.2	3247.55	❌
32	0048-PIHNL	Female	0	✅	❌	49	✅	No	No	No internet service	No internet service	No internet service	No internet service	No internet service	No internet service	One year	❌	Bank transfer (automatic)	20.45	900.9	❌
33	0052-DCKON	Male	0	✅	❌	66	✅	Yes	Fiber optic	Yes	Yes	Yes	Yes	Yes	Yes	One year	❌	Bank transfer (automatic)	115.8	7942.15	❌
34	0052-YNYOT	Female	0	❌	❌	67	✅	No	No	No internet service	No internet service	No internet service	No internet service	No internet service	No internet service	One year	❌	Electronic check	20.55	1343.4	❌
35	0056-EPFBG	Male	0	✅	✅	20	❌	No phone service	DSL	Yes	No	Yes	Yes	No	No	Two year	✅	Credit card (automatic)	39.4	825.4	❌
36	0057-QBUQH	Female	0	❌	✅	43	✅	Yes	No	No internet service	No internet service	No internet service	No internet service	No internet service	No internet service	Two year	✅	Electronic check	25.1	1070.15	❌
37	0058-EVZWM	Female	0	✅	❌	55	✅	Yes	Fiber optic	Yes	No	No	No	Yes	No	Month-to-month	✅	Bank transfer (automatic)	89.8	4959.6	❌
38	0060-FUALY	Female	0	✅	❌	59	✅	Yes	Fiber optic	Yes	Yes	No	No	Yes	No	Month-to-month	✅	Electronic check	94.75	5597.65	❌
39	0064-SUDOG	Female	0	✅	✅	12	✅	No	No	No internet service	No internet service	No internet service	No internet service	No internet service	No internet service	Two year	❌	Bank transfer (automatic)	20.3	224.5	❌
40	0064-YIJGF	Male	0	✅	✅	27	✅	Yes	Fiber optic	No	No	No	No	No	No	Month-to-month	✅	Bank transfer (automatic)	75.75	1929.0	❌
41	0067-DKWBL	Male	1	❌	❌	2	✅	No	DSL	Yes	No	No	No	No	No	Month-to-month	✅	Electronic check	49.25	91.1	✅
42	0068-FIGTF	Female	0	❌	❌	27	✅	No	DSL	No	Yes	Yes	Yes	Yes	Yes	One year	❌	Mailed check	78.2	2078.95	❌
43	0071-NDAFP	Male	0	✅	✅	25	✅	Yes	No	No internet service	No internet service	No internet service	No internet service	No internet service	No internet service	Two year	❌	Credit card (automatic)	25.5	630.6	❌
44	0074-HDKDG	Male	0	✅	✅	25	✅	No	DSL	Yes	Yes	Yes	No	No	No	One year	✅	Bank transfer (automatic)	61.6	1611.0	❌
45	0076-LVEPS	Male	0	❌	✅	29	❌	No phone service	DSL	Yes	Yes	Yes	Yes	No	No	Month-to-month	✅	Mailed check	45.0	1242.45	❌
46	0078-XZMHT	Male	0	✅	❌	72	✅	Yes	DSL	No	Yes	Yes	Yes	Yes	Yes	Two year	✅	Bank transfer (automatic)	85.15	6316.2	❌
47	0080-EMYVY	Female	0	❌	❌	14	✅	No	DSL	No	Yes	No	No	No	No	One year	❌	Credit card (automatic)	51.45	727.85	❌
48	0080-OROZO	Female	0	❌	❌	35	✅	No	Fiber optic	No	No	Yes	Yes	Yes	Yes	One year	✅	Electronic check	99.25	3532.0	❌
49	0082-LDZUE	Male	0	❌	❌	1	✅	No	DSL	No	No	No	No	No	No	Month-to-month	✅	Mailed check	44.3	44.3	❌
50	0082-OQIQY	Male	0	❌	❌	29	✅	No	Fiber optic	No	No	No	Yes	Yes	Yes	Month-to-month	✅	Electronic check	94.2	2607.6	❌
51	0083-PIVIK	Male	0	❌	❌	64	✅	Yes	DSL	Yes	Yes	Yes	Yes	Yes	No	One year	❌	Electronic check	81.25	5567.55	❌
52	0089-IIQKO	Female	0	✅	✅	39	✅	Yes	Fiber optic	No	No	No	Yes	Yes	Yes	Month-to-month	✅	Credit card (automatic)	99.95	3767.4	❌
53	0093-EXYQL	Female	1	❌	❌	40	✅	No	Fiber optic	No	No	No	No	Yes	Yes	Month-to-month	✅	Electronic check	91.55	3673.6	❌
54	0093-XWZFY	Male	0	❌	❌	40	✅	Yes	Fiber optic	No	Yes	Yes	No	Yes	Yes	Month-to-month	✅	Credit card (automatic)	104.5	4036.85	✅
55	0094-OIFMO	Female	1	❌	❌	11	✅	No	Fiber optic	No	Yes	No	No	Yes	Yes	Month-to-month	✅	Electronic check	95.0	1120.3	✅
56	0096-BXERS	Female	0	✅	❌	6	✅	Yes	DSL	No	No	No	No	No	No	Month-to-month	❌	Electronic check	50.35	314.55	❌
57	0096-FCPUF	Male	0	❌	❌	30	✅	Yes	DSL	Yes	No	No	No	No	Yes	Month-to-month	✅	Mailed check	64.5	1888.45	❌
58	0098-BOWSO	Male	0	❌	❌	27	✅	No	No	No internet service	No internet service	No internet service	No internet service	No internet service	No internet service	Month-to-month	✅	Electronic check	19.4	529.8	❌
59	0100-DUVFC	Male	1	✅	❌	70	✅	Yes	Fiber optic	No	Yes	Yes	No	Yes	Yes	One year	✅	Electronic check	104.8	7308.95	❌
60	0103-CSITQ	Female	0	✅	❌	57	✅	Yes	Fiber optic	No	Yes	Yes	Yes	Yes	Yes	One year	✅	Bank transfer (automatic)	109.4	6252.7	❌
61	0104-PPXDV	Male	0	✅	❌	58	✅	No	DSL	No	No	Yes	No	No	No	One year	❌	Credit card (automatic)	50.3	2878.55	❌
62	0106-GHRQR	Male	0	❌	❌	16	✅	Yes	DSL	No	Yes	Yes	No	No	Yes	Month-to-month	✅	Bank transfer (automatic)	71.4	1212.1	❌
63	0106-UGRDO	Female	0	✅	❌	69	✅	Yes	Fiber optic	Yes	Yes	Yes	Yes	Yes	Yes	Two year	✅	Electronic check	116.0	8182.85	❌
64	0107-WESLM	Male	0	❌	❌	1	✅	No	No	No internet service	No internet service	No internet service	No internet service	No internet service	No internet service	Month-to-month	✅	Electronic check	19.85	19.85	✅
65	0107-YHINA	Male	0	❌	✅	1	✅	Yes	Fiber optic	No	No	Yes	No	Yes	Yes	Month-to-month	✅	Electronic check	99.75	99.75	✅
66	0111-KLBQG	Male	1	✅	✅	32	✅	No	Fiber optic	No	Yes	No	No	Yes	Yes	Month-to-month	✅	Mailed check	93.95	2861.45	❌
67	0112-QAWRZ	Male	0	✅	✅	16	✅	Yes	Fiber optic	No	No	Yes	No	Yes	No	Month-to-month	✅	Bank transfer (automatic)	90.8	1442.2	❌
68	0112-QWPNC	Male	0	✅	❌	49	✅	Yes	DSL	Yes	No	Yes	Yes	Yes	Yes	One year	❌	Electronic check	84.35	4059.35	✅
69	0114-IGABW	Female	0	✅	❌	71	❌	No phone service	DSL	No	Yes	Yes	Yes	Yes	Yes	Two year	❌	Bank transfer (automatic)	58.25	4145.9	❌
70	0114-PEGZZ	Female	0	❌	❌	33	✅	Yes	Fiber optic	No	Yes	Yes	No	Yes	Yes	Month-to-month	❌	Electronic check	107.55	3645.5	❌
71	0114-RSRRW	Female	0	✅	❌	10	✅	No	No	No internet service	No internet service	No internet service	No internet service	No internet service	No internet service	Month-to-month	✅	Bank transfer (automatic)	19.95	187.75	❌
72	0115-TFERT	Male	0	✅	❌	21	✅	Yes	Fiber optic	No	Yes	Yes	Yes	Yes	Yes	Month-to-month	✅	Electronic check	111.2	2317.1	✅
73	0117-LFRMW	Male	0	✅	✅	37	❌	No phone service	DSL	Yes	Yes	Yes	No	No	No	Month-to-month	❌	Bank transfer (automatic)	40.2	1448.8	✅
74	0118-JPNOY	Female	1	❌	❌	26	✅	No	Fiber optic	Yes	No	No	No	Yes	No	Month-to-month	✅	Credit card (automatic)	85.8	2193.65	❌
75	0121-SNYRK	Male	0	❌	❌	50	❌	No phone service	DSL	Yes	No	No	Yes	No	No	One year	✅	Mailed check	35.4	1748.9	❌
76	0122-OAHPZ	Female	0	❌	❌	7	✅	Yes	Fiber optic	No	No	No	No	No	No	Month-to-month	✅	Electronic check	73.85	511.25	✅
77	0123-CRBRT	Female	0	✅	✅	61	✅	Yes	DSL	Yes	Yes	Yes	Yes	Yes	Yes	Two year	❌	Mailed check	88.1	5526.75	❌
78	0125-LZQXK	Male	0	❌	❌	15	✅	No	Fiber optic	No	No	Yes	Yes	Yes	Yes	Month-to-month	✅	Electronic check	101.35	1553.95	✅
79	0128-MKWSG	Female	0	❌	✅	26	❌	No phone service	DSL	Yes	No	No	Yes	No	Yes	Month-to-month	❌	Mailed check	45.8	1147.0	❌
80	0129-KPTWJ	Male	0	✅	❌	72	✅	No	Fiber optic	No	No	Yes	No	Yes	Yes	Month-to-month	✅	Electronic check	94.65	6747.35	❌
81	0129-QMPDR	Male	0	✅	✅	44	✅	No	No	No internet service	No internet service	No internet service	No internet service	No internet service	No internet service	One year	✅	Bank transfer (automatic)	20.5	865.05	❌
82	0130-SXOUN	Male	0	❌	❌	66	✅	Yes	Fiber optic	No	Yes	No	No	No	Yes	Month-to-month	❌	Credit card (automatic)	89.4	5976.9	❌
83	0133-BMFZO	Female	0	❌	❌	2	✅	No	Fiber optic	Yes	Yes	No	Yes	No	No	Month-to-month	✅	Electronic check	86.25	181.65	✅
84	0134-XWXCE	Female	1	❌	❌	44	✅	No	DSL	No	No	Yes	Yes	Yes	Yes	One year	❌	Bank transfer (automatic)	74.85	3268.05	❌
85	0135-NMXAP	Female	0	❌	❌	12	✅	Yes	Fiber optic	No	Yes	No	No	Yes	No	Month-to-month	✅	Bank transfer (automatic)	89.75	1052.4	✅
86	0136-IFMYD	Male	1	✅	❌	69	✅	Yes	Fiber optic	No	Yes	Yes	Yes	Yes	Yes	Two year	✅	Electronic check	109.95	7634.25	❌
87	0137-OCGAB	Female	0	❌	❌	1	✅	Yes	Fiber optic	No	Yes	No	No	No	No	Month-to-month	✅	Mailed check	80.2	80.2	✅
88	0137-UDEUO	Female	0	❌	❌	3	✅	No	No	No internet service	No internet service	No internet service	No internet service	No internet service	No internet service	Month-to-month	❌	Mailed check	19.85	63.75	✅
89	0139-IVFJG	Female	0	✅	❌	2	✅	No	Fiber optic	Yes	Yes	No	No	Yes	No	Month-to-month	❌	Electronic check	90.35	190.5	❌
90	0141-YEAYS	Female	1	❌	❌	27	✅	Yes	Fiber optic	No	Yes	Yes	No	No	No	Month-to-month	✅	Bank transfer (automatic)	86.45	2401.05	❌
91	0142-GVYSN	Male	0	❌	❌	26	✅	No	No	No internet service	No internet service	No internet service	No internet service	No internet service	No internet service	Month-to-month	✅	Electronic check	20.3	511.25	❌
92	0147-ESWWR	Female	1	✅	❌	39	✅	Yes	Fiber optic	No	No	No	Yes	Yes	Yes	Month-to-month	✅	Electronic check	101.25	3949.15	❌
93	0148-DCDOS	Male	0	❌	❌	25	✅	Yes	Fiber optic	No	No	No	No	Yes	Yes	Month-to-month	✅	Bank transfer (automatic)	94.7	2362.1	✅
94	0151-ONTOV	Female	0	❌	❌	1	✅	No	Fiber optic	No	No	No	No	No	No	Month-to-month	❌	Mailed check	70.9	70.9	✅
95	0156-FVPTA	Male	0	✅	❌	22	✅	No	DSL	Yes	No	No	Yes	No	No	Month-to-month	✅	Electronic check	54.2	1152.7	✅
96	0164-APGRB	Female	0	❌	❌	72	✅	Yes	Fiber optic	Yes	Yes	Yes	Yes	Yes	Yes	Two year	✅	Electronic check	114.9	8496.7	❌
97	0164-XAIRP	Female	0	❌	❌	24	✅	No	No	No internet service	No internet service	No internet service	No internet service	No internet service	No internet service	Two year	✅	Bank transfer (automatic)	19.55	470.2	❌
98	0168-XZKBB	Female	0	✅	❌	19	✅	Yes	Fiber optic	No	No	No	No	Yes	No	Month-to-month	✅	Electronic check	86.85	1564.4	❌
99	0174-QRVVY	Male	0	✅	✅	71	✅	Yes	No	No internet service	No internet service	No internet service	No internet service	No internet service	No internet service	Two year	❌	Credit card (automatic)	25.35	1847.55	❌
100	0177-PXBAT	Male	1	✅	❌	33	✅	Yes	Fiber optic	Yes	No	Yes	Yes	Yes	Yes	Month-to-month	✅	Bank transfer (automatic)	109.9	3694.7	❌

Rows: 1-100 | Columns: 21

The Pearson correlation coefficient is a very common correlation function. In this case, it helped us to find linear links between the variables. Having a strong Pearson relationship means that the two input variables are linearly correlated.

In [2]:

vdf.corr(method = "pearson")

Out[2]:

	"SeniorCitizen"	"Partner"	"Dependents"	"tenure"	"PhoneService"	"PaperlessBilling"	"MonthlyCharges"	"TotalCharges"	"Churn"
"SeniorCitizen"	1.0	0.0164786575974139	-0.211185088493958	0.0165668776818093	0.00857640107927944	0.156529559311173	0.220173338576274	0.102410605395332	0.150889328176473
"Partner"	0.0164786575974139	1.0	0.452676282929464	0.379697461168293	0.017705663223972	-0.014876622287891	0.096847936256471	0.319072363238569	-0.150447544959177
"Dependents"	-0.211185088493958	0.452676282929464	1.0	0.159712331216918	-0.00176167854468371	-0.111377229193644	-0.113890230021937	0.0646532494217718	-0.164221401579725
"tenure"	0.0165668776818093	0.379697461168293	0.159712331216918	1.0	0.00844820673274375	0.00615248159827112	0.24789985628615	0.825880460933193	-0.352228670113078
"PhoneService"	0.00857640107927944	0.017705663223972	-0.00176167854468371	0.00844820673274375	1.0	0.0165048057325697	0.24739796321414	0.113008260954741	0.0119419800290031
"PaperlessBilling"	0.156529559311173	-0.014876622287891	-0.111377229193644	0.00615248159827112	0.0165048057325697	1.0	0.352149967795987	0.157829782865907	0.191825331666468
"MonthlyCharges"	0.220173338576274	0.096847936256471	-0.113890230021937	0.24789985628615	0.24739796321414	0.352149967795987	1.0	0.651064803226215	0.193356422237847
"TotalCharges"	0.102410605395332	0.319072363238569	0.0646532494217718	0.825880460933193	0.113008260954741	0.157829782865907	0.651064803226215	1.0	-0.199484083567563
"Churn"	0.150889328176473	-0.150447544959177	-0.164221401579725	-0.352228670113078	0.0119419800290031	0.191825331666468	0.193356422237847	-0.199484083567563	1.0

Rows: 1-9 | Columns: 10

We can see that 'tenure' is well-correlated to the 'TotalCharges', which makes sense.

In [3]:

vdf.scatter(["tenure", "TotalCharges"])
vdf.corr(["tenure", "TotalCharges"], method = "pearson")

Out[3]:

0.825880460933193

Note, however, that having a low Pearson relationship imply that the variables aren't correlated. For example, let's compute the Pearson correlation coefficient between 'tenure' and 'TotalCharges' to the power of 20.

In [4]:

vdf["TotalCharges^20"] = vdf["TotalCharges"] ** 20
vdf.scatter(["tenure", "TotalCharges^20"])
vdf.corr(["tenure", "TotalCharges^20"], method = "pearson")

Out[4]:

0.224994408804537

We know that the 'tenure' and 'TotalCharges' are strongly linearly correlated. However we can notice that the correlation between the 'tenure' and 'TotalCharges' to the power of 20 is not very high. Indeed, the Pearson correlation coefficient is not robust for monotonic relationships, but rank-based correlations are. Knowing this, we'll calculate the Spearman's rank correlation coefficient instead.

In [5]:

vdf.corr(method = "spearman")

Out[5]:

	"SeniorCitizen"	"Partner"	"Dependents"	"tenure"	"PhoneService"	"PaperlessBilling"	"MonthlyCharges"	"TotalCharges"	"Churn"	"TotalCharges^20"
"SeniorCitizen"	1.0	0.0164786575974139	-0.211185088493958	0.0190767898701152	0.00857640107927944	0.156529559311173	0.221092529102162	0.105795342303725	0.150889328176473	0.105795342303725
"Partner"	0.0164786575974139	1.0	0.452676282929464	0.384665710284119	0.017705663223972	-0.014876622287891	0.108410945895981	0.343930553215626	-0.150447544959177	0.343930553215626
"Dependents"	-0.211185088493958	0.452676282929464	1.0	0.164485741353804	-0.00176167854468371	-0.111377229193644	-0.107082725586711	0.0866797760484616	-0.164221401579725	0.0866797760484616
"tenure"	0.0190767898701152	0.384665710284119	0.164485741353804	1.0	0.00815081986907184	0.00792876239476321	0.276342245223708	0.883103368818293	-0.369620778763435	0.883103368818293
"PhoneService"	0.00857640107927944	0.017705663223972	-0.00176167854468371	0.00815081986907184	1.0	0.0165048057325697	0.238826410230016	0.0838048547856037	0.0119419800290031	0.0838048547856037
"PaperlessBilling"	0.156529559311173	-0.014876622287891	-0.111377229193644	0.00792876239476321	0.0165048057325697	1.0	0.346158879381323	0.151669712799097	0.191825331666468	0.151669712799097
"MonthlyCharges"	0.221092529102162	0.108410945895981	-0.107082725586711	0.276342245223708	0.238826410230016	0.346158879381323	1.0	0.633958405301206	0.184839285783758	0.633958405301206
"TotalCharges"	0.105795342303725	0.343930553215626	0.0866797760484616	0.883103368818293	0.0838048547856037	0.151669712799097	0.633958405301206	1.0	-0.233211018585104	1.0
"Churn"	0.150889328176473	-0.150447544959177	-0.164221401579725	-0.369620778763435	0.0119419800290031	0.191825331666468	0.184839285783758	-0.233211018585104	1.0	-0.233211018585104
"TotalCharges^20"	0.105795342303725	0.343930553215626	0.0866797760484616	0.883103368818293	0.0838048547856037	0.151669712799097	0.633958405301206	1.0	-0.233211018585104	1.0

Rows: 1-10 | Columns: 11

The Spearman's rank correlation coefficient determines the monotonic relationships between the variables.

In [6]:

vdf.corr(["tenure", "TotalCharges^20"], method = "spearman")

Out[6]:

0.883103368818293

We can notice that Spearman's rank correlation coefficient stays the same if one of the variables can be expressed using a monotonic function on the other. The same applies to Kendall rank correlation coefficient.

In [7]:

vdf.corr(method = "kendall")

Out[7]:

	"SeniorCitizen"	"Partner"	"Dependents"	"tenure"	"PhoneService"	"PaperlessBilling"	"MonthlyCharges"	"TotalCharges"	"Churn"	"TotalCharges^20"
"SeniorCitizen"	1	0.0164786575974139	-0.211185088493958	0.0153154063543276	0.00857640107927944	0.156529559311173	0.180635949216048	0.0873571166085645	0.150889328176473	0.0873571166085645
"Partner"	0.0164786575974139	1	0.452676282929464	0.317713503646956	0.017705663223972	-0.014876622287891	0.0886260353369454	0.279358144937778	-0.150447544959177	0.279358144937778
"Dependents"	-0.211185088493958	0.452676282929464	1	0.135506499930892	-0.00176167854468371	-0.111377229193644	-0.0874394375642774	0.0674036111339054	-0.164221401579725	0.0674036111339054
"tenure"	0.0153154063543276	0.317713503646956	0.135506499930892	1	0.00700180312561143	0.0063514568219607	0.190564868116964	0.731699318287362	-0.302965018211126	0.731699318287362
"PhoneService"	0.00857640107927944	0.017705663223972	-0.00176167854468371	0.00700180312561143	1	0.0165048057325697	0.195862538105968	0.0690761344143585	0.0119419800290031	0.0690761344143585
"PaperlessBilling"	0.156529559311173	-0.014876622287891	-0.111377229193644	0.0063514568219607	0.0165048057325697	1	0.28271562697138	0.12528100183927	0.191825331666468	0.12528100183927
"MonthlyCharges"	0.180635949216048	0.0886260353369454	-0.0874394375642774	0.190564868116964	0.195862538105968	0.28271562697138	1	0.458469198925234	0.150955285158245	0.458469198925234
"TotalCharges"	0.0873571166085645	0.279358144937778	0.0674036111339054	0.731699318287362	0.0690761344143585	0.12528100183927	0.458469198925234	1	-0.189101699239313	0.996878225678041
"Churn"	0.150889328176473	-0.150447544959177	-0.164221401579725	-0.302965018211126	0.0119419800290031	0.191825331666468	0.150955285158245	-0.189101699239313	1	-0.189101699239313
"TotalCharges^20"	0.0873571166085645	0.279358144937778	0.0674036111339054	0.731699318287362	0.0690761344143585	0.12528100183927	0.458469198925234	0.996878225678041	-0.189101699239313	1

Rows: 1-10 | Columns: 11

Notice that the Kendall rank correlation coefficient will also detect the monotonic relationship.

In [8]:

vdf.corr(["tenure", "TotalCharges^20"], method = "kendall")

Out[8]:

0.731699318287362

However, the Kendall rank correlation coefficient is very computationally expensive, so we'll generally use Pearson and Spearman when dealing with correlations between numerical variables.

Binary features are considered numerical, but this isn't technically accurate. Since binary variables can only take two values, calculating correlations between a binary and numerical variable can lead to misleading results. To account for this, we'll want to use the 'Biserial Point' method to calculate the Point-Biserial correlation coefficient. This powerful method will help us understand the link between a binary variable and a numerical variable.

In [9]:

vdf.corr(method = "biserial")

Out[9]:

	"SeniorCitizen"	"Partner"	"Dependents"	"tenure"	"PhoneService"	"PaperlessBilling"	"MonthlyCharges"	"TotalCharges"	"Churn"	"TotalCharges^20"
"SeniorCitizen"	1	0.0164774876952011	-0.211170095409668	0.0165657015164121	0.0085757921977281	0.156518446495901	0.220157707371214	0.102403323381215	0.150878615789382	0.00959483426594477
"Partner"	0.0164774876952011	1	0.452644145179047	0.379670504548958	0.017704406210501	-0.014875566122167	0.09684106053773	0.319049675261633	-0.150436863936483	0.0945722425175577
"Dependents"	-0.211170095409668	0.452644145179047	1	0.159700992440771	-0.0017615534743243	-0.111369321967821	-0.113882144385631	0.0646486521842783	-0.164209742682151	0.0224428704447468
"tenure"	0.0165657015164121	0.379670504548958	0.159700992440771	1	nan	nan	nan	nan	nan	nan
"PhoneService"	0.0085757921977281	0.017704406210501	-0.0017615534743243	nan	1	0.0165036339739701	0.247380399196989	0.113000225383144	0.0119411322081913	0.0467146740346801
"PaperlessBilling"	0.156518446495901	-0.014875566122167	-0.111369321967821	nan	0.0165036339739701	1	0.352124966910804	0.157818560212722	0.191811713028016	0.0579931118233683
"MonthlyCharges"	0.220157707371214	0.09684106053773	-0.113882144385631	nan	0.247380399196989	0.352124966910804	1	nan	0.193342694899628	nan
"TotalCharges"	0.102403323381215	0.319049675261633	0.0646486521842783	nan	0.113000225383144	0.157818560212722	nan	1	-0.199469899041392	nan
"Churn"	0.150878615789382	-0.150436863936483	-0.164209742682151	nan	0.0119411322081913	0.191811713028016	0.193342694899628	-0.199469899041392	1	-0.0630928693846116
"TotalCharges^20"	0.00959483426594477	0.0945722425175577	0.0224428704447468	nan	0.0467146740346801	0.0579931118233683	nan	nan	-0.0630928693846116	1

Rows: 1-10 | Columns: 11

Lastly, we'll look at the relationship between categorical columns. In this case, the 'Cramer's V' method is very efficient. Since there is no position in the Euclidean space for those variables, the 'Cramer's V' coefficients cannot be negative (which is a sign of an opposite relationship) and they will range in the interval [0,1].

In [10]:

vdf.corr(method = "cramer")

Out[10]:

	"customerID"	"gender"	"SeniorCitizen"	"Partner"	"Dependents"	"PhoneService"	"MultipleLines"	"InternetService"	"OnlineSecurity"	"OnlineBackup"	"DeviceProtection"	"TechSupport"	"StreamingTV"	"StreamingMovies"	"Contract"	"PaperlessBilling"	"PaymentMethod"	"Churn"
"customerID"	1	0.7071387759872941	0.8533988239339025	0.7075137962272524	0.7617937674788702	0.9083424868747245	0.8423185423961393	0.8241182589052664	0.8293968523170167	0.824052991235141	0.8241100151107151	0.8289549641545605	0.822756376161546	0.8227324857138872	0.8379652208984673	0.7190332195604549	0.8679291439283205	0.7810906445878953
"gender"	0.7071387759872941	1	0.001873711638260736	0.001808079661789021	0.010516611922734077	0.006487672136208199	0.009451666570452976	0.011319816023820888	0.0170248673996952	0.013798357377745577	0.0060292442384310314	0.009722241217455355	0.008844328042668036	0.010640764864840464	0.008152557349604891	0.011753691221222193	0.01864584107758851	0.008612095078997812
"SeniorCitizen"	0.8533988239339025	0.001873711638260736	1	0.016478657597413914	0.21118508849395765	0.008576401079279436	0.1466428981547604	0.26532401247490706	0.21107026551439387	0.18277371600205183	0.18307117871498424	0.22405049807030314	0.18558579209841322	0.18896193131088726	0.1438683664984389	0.15652955931117277	0.19598451589487909	0.1508893281764731
"Partner"	0.7075137962272524	0.001808079661789021	0.016478657597413914	1	0.4526762829294638	0.017705663223972032	0.14389768511332032	0.0009073521208289714	0.1520098144303294	0.15333685221744686	0.16656826190673807	0.12764156082912625	0.13734265561456047	0.1295738983957767	0.29603947593743685	0.01487662228789104	0.16276227893705197	0.15044754495917656
"Dependents"	0.7617937674788702	0.010516611922734077	0.21118508849395765	0.4526762829294638	1	0.0017616785446837105	0.025094016713979587	0.1802672478825538	0.19457924376849997	0.1627601413923546	0.15757010710298983	0.1823774404156106	0.14708243957321443	0.1413353739608838	0.24421223608441336	0.11137722919364426	0.15074411049283012	0.16422140157972528
"PhoneService"	0.9083424868747245	0.006487672136208199	0.008576401079279436	0.017705663223972032	0.0017616785446837105	1	0.9083424868747245	0.3756877098193481	0.1004543340659466	0.09426132351692645	0.0933403708615453	0.10168502026979014	0.1075905732294665	0.10262843139014985	0.003976117347714681	0.01650480573256966	0.010297221766590077	0.01194198002900308
"MultipleLines"	0.8423185423961393	0.009451666570452976	0.1466428981547604	0.14389768511332032	0.025094016713979587	0.9083424868747245	1	0.35442564054609377	0.20549824502887423	0.22274250160703957	0.22404038042422336	0.20611678556080693	0.24092390664877442	0.24153764189544388	0.08040766226914967	0.16634063233977742	0.1669298080382473	0.04010925283378872
"InternetService"	0.8241182589052664	0.011319816023820888	0.26532401247490706	0.0009073521208289714	0.1802672478825538	0.3756877098193481	0.35442564054609377	1	0.5959261119397458	0.5747918951855748	0.5746979259753913	0.5940263647919771	0.5869475572452835	0.5856137246362116	0.20692091402897067	0.3784273654205174	0.3131230023778106	0.32245455521230887
"OnlineSecurity"	0.8293968523170167	0.0170248673996952	0.21107026551439387	0.1520098144303294	0.19457924376849997	0.1004543340659466	0.20549824502887423	0.5959261119397458	1	0.5885783134226351	0.587182704257195	0.6063578188342443	0.575535268964284	0.5760517524952443	0.30134873250567035	0.34179975542511276	0.30495874601380435	0.3474004326740551
"OnlineBackup"	0.824052991235141	0.013798357377745577	0.18277371600205183	0.15333685221744686	0.1627601413923546	0.09426132351692645	0.22274250160703957	0.5747918951855748	0.5885783134226351	1	0.5894213880027672	0.5902863075237831	0.5840131835568888	0.5827676397944742	0.2585717875284732	0.3210458207980537	0.282516640706162	0.2923155121954445
"DeviceProtection"	0.8241100151107151	0.0060292442384310314	0.18307117871498424	0.16656826190673807	0.15757010710298983	0.0933403708615453	0.22404038042422336	0.5746979259753913	0.587182704257195	0.5894213880027672	1	0.5983789246078595	0.6070707969365265	0.6099527291303553	0.29697649313511804	0.3216312457665022	0.28702116041733605	0.281579732968073
"TechSupport"	0.8289549641545605	0.009722241217455355	0.22405049807030314	0.12764156082912625	0.1823774404156106	0.10168502026979014	0.20611678556080693	0.5940263647919771	0.6063578188342443	0.5902863075237831	0.5983789246078595	1	0.5859300440849506	0.5860036200553587	0.3310209176488737	0.3295243087543918	0.3068883711219149	0.3429161982469257
"StreamingTV"	0.822756376161546	0.008844328042668036	0.18558579209841322	0.13734265561456047	0.14708243957321443	0.1075905732294665	0.24092390664877442	0.5869475572452835	0.575535268964284	0.5840131835568888	0.6070707969365265	0.5859300440849506	1	0.6517530725113684	0.2359013654392877	0.3360606862742847	0.2736710577877493	0.2305023384466809
"StreamingMovies"	0.8227324857138872	0.010640764864840464	0.18896193131088726	0.1295738983957767	0.1413353739608838	0.10262843139014985	0.24153764189544388	0.5856137246362116	0.5760517524952443	0.5827676397944742	0.6099527291303553	0.5860036200553587	0.6517530725113684	1	0.238639856353762	0.3321109249335917	0.27465873933463103	0.230950809069268
"Contract"	0.8379652208984673	0.008152557349604891	0.1438683664984389	0.29603947593743685	0.24421223608441336	0.003976117347714681	0.08040766226914967	0.20692091402897067	0.30134873250567035	0.2585717875284732	0.29697649313511804	0.3310209176488737	0.2359013654392877	0.238639856353762	1	0.17764521997131164	0.2666547476541491	0.4101156965761409
"PaperlessBilling"	0.7190332195604549	0.011753691221222193	0.15652955931117277	0.01487662228789104	0.11137722919364426	0.01650480573256966	0.16634063233977742	0.3784273654205174	0.34179975542511276	0.3210458207980537	0.3216312457665022	0.3295243087543918	0.3360606862742847	0.3321109249335917	0.17764521997131164	1	0.24871857267198466	0.1918253316664679
"PaymentMethod"	0.8679291439283205	0.01864584107758851	0.19598451589487909	0.16276227893705197	0.15074411049283012	0.010297221766590077	0.1669298080382473	0.3131230023778106	0.30495874601380435	0.282516640706162	0.28702116041733605	0.3068883711219149	0.2736710577877493	0.27465873933463103	0.2666547476541491	0.24871857267198466	1	0.3033586255540706
"Churn"	0.7810906445878953	0.008612095078997812	0.1508893281764731	0.15044754495917656	0.16422140157972528	0.01194198002900308	0.04010925283378872	0.32245455521230887	0.3474004326740551	0.2923155121954445	0.281579732968073	0.3429161982469257	0.2305023384466809	0.230950809069268	0.4101156965761409	0.1918253316664679	0.3033586255540706	1

Rows: 1-18 | Columns: 19

Sometimes, we just need to look at the correlation between a response and other variables. The parameter 'focus' will isolate and show us the specified correlation vector.

In [11]:

vdf.corr(method = "cramer", focus = "Churn")

Out[11]:

	"Churn"
"Churn"	1
"customerID"	0.7810906445878953
"Contract"	0.4101156965761409
"OnlineSecurity"	0.3474004326740551
"TechSupport"	0.3429161982469257
"InternetService"	0.32245455521230887
"PaymentMethod"	0.3033586255540706
"OnlineBackup"	0.2923155121954445
"DeviceProtection"	0.281579732968073
"StreamingMovies"	0.230950809069268
"StreamingTV"	0.2305023384466809
"PaperlessBilling"	0.1918253316664679
"Dependents"	0.16422140157972528
"SeniorCitizen"	0.1508893281764731
"Partner"	0.15044754495917656
"MultipleLines"	0.04010925283378872
"PhoneService"	0.01194198002900308
"gender"	0.008612095078997812

Rows: 1-18 | Columns: 2

Sometimes a correlation coefficient can lead to incorrect assumptions, so we should always look at the coefficient p-value.

In [12]:

vdf.corr_pvalue("Churn", "customerID", method = "cramer",)

Out[12]:

(0.7810906445878953, 1.3659871749110484e-36)

We can see that churning correlates to the type of contract (monthly, yearly, etc.) which makes sense: you would expect that different types of contracts differ in flexibility for the customer, and particularly restrictive contracts may make churning more likely.

The type of internet service also seems to correlate with churning. Let's split the different categories to binaries to understand which services can influence the global churning rate.

In [13]:

vdf["InternetService"].one_hot_encode()
vdf.corr(method = "spearman", 
         focus = "Churn", 
         columns = ["InternetService_DSL", 
                    "InternetService_Fiber_optic"])

Out[13]:

	"Churn"
"InternetService_Fiber_optic"	0.308
"InternetService_DSL"	-0.124

Rows: 1-2 | Columns: 2

We can see that the Fiber Optic option in particular seems to be directly linked to a customer's likelihood to churn. Let's compute some aggregations to find a causal relationship.

In [14]:

vdf["contract"].one_hot_encode()
vdf.groupby(["InternetService_Fiber_optic"], 
            ["AVG(tenure) AS tenure", 
             "AVG(totalcharges) AS totalcharges",
             'AVG("contract_month-to-month") AS "contract_month-to-month"',
             'AVG("monthlycharges") AS "monthlycharges"'])

Out[14]:

	123 InternetService_Fiber_optic Integer	123 tenure Float	123 totalcharges Float	123 contract_month-to-month Float	123 monthlycharges Float
1	0	31.9422346085635	1558.06548526423	0.442614644033443	43.7882442361287
2	1	32.9179586563307	3205.30457041344	0.68733850129199	91.5001291989664

Rows: 1-2 | Columns: 5

It seems that users with the Fiber Optic option tend more to churn not because of the option itself, but probably because of the type of contracts and the monthly charges the users are paying to get it. Be careful when dealing with identifying correlations! Remember: correlation doesn't imply causation!

Another important type of correlation is the autocorrelation. Let's use the Amazon dataset to understand it.

In [15]:

from verticapy.datasets import load_amazon
vdf = load_amazon()
display(vdf)

	📅 date Date	Abc state Varchar(32)	123 number Int
1	1998-01-01	ACRE	0
2	1998-01-01	ALAGOAS	0
3	1998-01-01	AMAPÁ	0
4	1998-01-01	AMAZONAS	0
5	1998-01-01	BAHIA	0
6	1998-01-01	CEARÁ	0
7	1998-01-01	DISTRITO FEDERAL	0
8	1998-01-01	ESPÍRITO SANTO	0
9	1998-01-01	GOIÁS	0
10	1998-01-01	MARANHÃO	0
11	1998-01-01	MATO GROSSO	0
12	1998-01-01	MATO GROSSO DO SUL	0
13	1998-01-01	MINAS GERAIS	0
14	1998-01-01	PARANÁ	0
15	1998-01-01	PARAÍBA	0
16	1998-01-01	PARÁ	0
17	1998-01-01	PERNAMBUCO	0
18	1998-01-01	PIAUÍ	0
19	1998-01-01	RIO DE JANEIRO	0
20	1998-01-01	RIO GRANDE DO NORTE	0
21	1998-01-01	RIO GRANDE DO SUL	0
22	1998-01-01	RONDÔNIA	0
23	1998-01-01	RORAIMA	0
24	1998-01-01	SANTA CATARINA	0
25	1998-01-01	SERGIPE	0
26	1998-01-01	SÃO PAULO	0
27	1998-01-01	TOCANTINS	0
28	1998-02-01	ACRE	0
29	1998-02-01	ALAGOAS	0
30	1998-02-01	AMAPÁ	0
31	1998-02-01	AMAZONAS	0
32	1998-02-01	BAHIA	0
33	1998-02-01	CEARÁ	0
34	1998-02-01	DISTRITO FEDERAL	0
35	1998-02-01	ESPÍRITO SANTO	0
36	1998-02-01	GOIÁS	0
37	1998-02-01	MARANHÃO	0
38	1998-02-01	MATO GROSSO	0
39	1998-02-01	MATO GROSSO DO SUL	0
40	1998-02-01	MINAS GERAIS	0
41	1998-02-01	PARANÁ	0
42	1998-02-01	PARAÍBA	0
43	1998-02-01	PARÁ	0
44	1998-02-01	PERNAMBUCO	0
45	1998-02-01	PIAUÍ	0
46	1998-02-01	RIO DE JANEIRO	0
47	1998-02-01	RIO GRANDE DO NORTE	0
48	1998-02-01	RIO GRANDE DO SUL	0
49	1998-02-01	RONDÔNIA	0
50	1998-02-01	RORAIMA	0
51	1998-02-01	SANTA CATARINA	0
52	1998-02-01	SERGIPE	0
53	1998-02-01	SÃO PAULO	0
54	1998-02-01	TOCANTINS	0
55	1998-03-01	ACRE	0
56	1998-03-01	ALAGOAS	0
57	1998-03-01	AMAPÁ	0
58	1998-03-01	AMAZONAS	0
59	1998-03-01	BAHIA	0
60	1998-03-01	CEARÁ	0
61	1998-03-01	DISTRITO FEDERAL	0
62	1998-03-01	ESPÍRITO SANTO	0
63	1998-03-01	GOIÁS	0
64	1998-03-01	MARANHÃO	0
65	1998-03-01	MATO GROSSO	0
66	1998-03-01	MATO GROSSO DO SUL	0
67	1998-03-01	MINAS GERAIS	0
68	1998-03-01	PARANÁ	0
69	1998-03-01	PARAÍBA	0
70	1998-03-01	PARÁ	0
71	1998-03-01	PERNAMBUCO	0
72	1998-03-01	PIAUÍ	0
73	1998-03-01	RIO DE JANEIRO	0
74	1998-03-01	RIO GRANDE DO NORTE	0
75	1998-03-01	RIO GRANDE DO SUL	0
76	1998-03-01	RONDÔNIA	0
77	1998-03-01	RORAIMA	0
78	1998-03-01	SANTA CATARINA	0
79	1998-03-01	SERGIPE	0
80	1998-03-01	SÃO PAULO	0
81	1998-03-01	TOCANTINS	0
82	1998-04-01	ACRE	0
83	1998-04-01	ALAGOAS	0
84	1998-04-01	AMAPÁ	0
85	1998-04-01	AMAZONAS	0
86	1998-04-01	BAHIA	0
87	1998-04-01	CEARÁ	0
88	1998-04-01	DISTRITO FEDERAL	0
89	1998-04-01	ESPÍRITO SANTO	0
90	1998-04-01	GOIÁS	0
91	1998-04-01	MARANHÃO	0
92	1998-04-01	MATO GROSSO	0
93	1998-04-01	MATO GROSSO DO SUL	0
94	1998-04-01	MINAS GERAIS	0
95	1998-04-01	PARANÁ	0
96	1998-04-01	PARAÍBA	0
97	1998-04-01	PARÁ	0
98	1998-04-01	PERNAMBUCO	0
99	1998-04-01	PIAUÍ	0
100	1998-04-01	RIO DE JANEIRO	0

Rows: 1-100 | Columns: 3

Our goal is to predict the number of forest fires in Brazil. To do this, we can draw an autocorrelation plot and a partial autocorrelation plot.

In [16]:

vdf.acf(column = "number",
        ts = "date",
        by = ["state"],
        p = 48,
        method = "pearson")
vdf.pacf(column = "number",
         ts = "date",
         by = ["state"],
         p = 48)

Out[16]:

	value	confidence
0	1.0	0.024396841824873755
1	0.672667529541858	0.033675228780048694
2	-0.188727403801382	0.034301750180211296
3	0.022206688265849	0.03431296790442547
4	-0.0819798501305434	0.03443207317154983
5	-0.00663606854011195	0.03443550440495167
6	-0.00273044444174287	0.03443830352079605
7	0.014509850994978	0.03444461651417873
8	0.0984841861349423	0.03461467835735325
9	0.144012733751383	0.034972633390635596
10	0.234587317293272	0.03590106333622817
11	0.424449663688596	0.03878033225579347
12	0.462136890390786	0.04193874171442886
13	-0.170929292614458	0.04235541732162943
14	-0.0309835645206131	0.0423722220603323
15	-0.0201661466233994	0.042381237262525503
16	0.00880380742618245	0.042385619747699776
17	-0.0144170836711788	0.04239183813723341
18	0.0105864910986754	0.04239670917644579
19	-3.04350920105641e-05	0.04240000328911371
20	0.0126700457251592	0.04240555842646121
21	0.061941763258756	0.042462844676922726
22	0.121170469484258	0.04267213673556238
23	0.24924674031484	0.04353632913341579
24	0.249959745256812	0.044388749935414466
25	-0.151776015949538	0.0447011914060506
26	-0.0432756193626391	0.044729696703441166
27	0.0425724972388961	0.04475738662368059
28	-0.051161634227943	0.04479581330579359
29	0.0223139130513636	0.0448059438601625
30	-0.0149213359411028	0.044812402245112425
31	-0.00291932296754487	0.04481600427562471
32	-0.00105989613246084	0.04481950838871394
33	0.0296382223991214	0.04483472136181398
34	0.00899438559236398	0.044839292596148154
35	0.211433234463727	0.04543546774174034
36	0.171820891345056	0.045826240829950876
37	-0.0701311831783696	0.045894011277228176
38	-0.0070002925494054	0.04589822691705164
39	0.0328060179320583	0.04591584235878527
40	-0.0426815603685585	0.045943175691876184
41	0.00149108886355715	0.04594678657275398
42	-0.00947695003968667	0.045951540276646934
43	-0.0208017880882059	0.04596076565630418
44	-0.0216657270948819	0.04597047036327498
45	-0.0313007366739144	0.045986828122287306
46	-0.0540812884438045	0.04602852459945253
47	-0.0230892672986662	0.04603905984310686
48	-0.0661604548541157	0.046099627206536

Rows: 1-49 | Columns: 3

We can see the seasonality forest fires.

It's mathematically impossible to build the perfect correlation function, but we still have several powerful functions at our disposal for finding relationships in all kinds of datasets.