CODEWITHSUNDEEP
rpercentage
R Maharaj
R Maharaj

Reputation: 115

Rolling week-on-week % change

I have some time series daily data as such:

        date  cases
1   5/1/2022  26582
2   5/2/2022  14921
3   5/3/2022  38811
4   5/4/2022  20838
5   5/5/2022  19382
6   5/6/2022  18768
7   5/7/2022  49466
8   5/8/2022  40081
9   5/9/2022  19993
10 5/10/2022  49662
11 5/11/2022  45268
12 5/12/2022  38406
13 5/13/2022  36880
14 5/14/2022  39393
15 5/15/2022  30418
16 5/16/2022  15220
17 5/17/2022  43684
18 5/18/2022  42053
19 5/19/2022  36577
20 5/20/2022  34825
21 5/21/2022  35802
22 5/22/2022  26678
23 5/23/2022  12743
24 5/24/2022  37721
25 5/25/2022  34616
26 5/26/2022  28079
27 5/27/2022  24891
28 5/28/2022  25026
29 5/29/2022  17369
30 5/30/2022   8367
31 5/31/2022  24818
32  6/1/2022  22847
33  6/2/2022  19334
34  6/3/2022  17917
35  6/4/2022  17927
36  6/5/2022  13350
37  6/6/2022   6589
38  6/7/2022  20085
39  6/8/2022  18234
40  6/9/2022  15891
41 6/10/2022  14707
42 6/11/2022  15422
43 6/12/2022  11845
44 6/13/2022   6001
45 6/14/2022  17742
46 6/15/2022  16497
47 6/16/2022  14833
48 6/17/2022  13838
49 6/18/2022  15260
50 6/19/2022  11630
51 6/20/2022   5720
52 6/21/2022  18223
53 6/22/2022  17182
54 6/23/2022  16043
55 6/24/2022  14879
56 6/25/2022  16492
57 6/26/2022  12995
58 6/27/2022   7140
59 6/28/2022  22927
60 6/29/2022  23376
61 6/30/2022  22515
62  7/1/2022  24113
63  7/2/2022  27026
64  7/3/2022  21813
65  7/4/2022  13364
66  7/5/2022  44746
67  7/6/2022  47535
68  7/7/2022  48275
69  7/8/2022  49235
70  7/9/2022  59011
71 7/10/2022  50908
72 7/11/2022  27452
73 7/12/2022  97771
74 7/13/2022 102621
75 7/14/2022  98980
76 7/15/2022  99763
77 7/16/2022 117180
78 7/17/2022  99985
79 7/18/2022  76164
80 7/19/2022  66725
81 7/20/2022 152495
82 7/21/2022 186209
83 7/22/2022 195136
84 7/23/2022 200937
85 7/24/2022 176517
86 7/25/2022 126535
87 7/26/2022 196453
88 7/27/2022 209657
89 7/28/2022 233066
90 7/29/2022 221409
91 7/30/2022 222274
92 7/31/2022 197750
93  8/1/2022 139668

I require to find the rolling week-on-week % change in cases, which is basically the % change in cases over the last 7 days relative to the previous 7 days. I know how to implement this on Excel but unsure how to do this on R. To illustrate what I require:

% change on 5/14/2022 = ([(sum of cases over from row 14 to 8) / (sum of cases from row 7 to 1)] - 1)*100

then % change on 5/15/2022 = ([(sum of cases over from row 15 to 9) / (sum of cases from row 8 to 2)] - 1)*100 and so on...

Any help will be greatly appreciated.

Upvotes: 0

Views: 72

Answers (1)

G. Grothendieck
G. Grothendieck

Reputation: 269481

Define function chg as the percent change given 14 days of data and use rollapplyr to apply it in a rolling fashion.

library(zoo)

chg <- function(x) 100 * sum(tail(x, 7)) / sum(head(x, 7)) - 100
transform(DF, chg = rollapplyr(cases, 14, chg, fill = NA))

Note

First 20 rows.

Lines <- "        date cases
1   5/1/2022 26582
2   5/2/2022 14921
3   5/3/2022 38811
4   5/4/2022 20838
5   5/5/2022 19382
6   5/6/2022 18768
7   5/7/2022 49466
8   5/8/2022 40081
9   5/9/2022 19993
10 5/10/2022 49662
11 5/11/2022 45268
12 5/12/2022 38406
13 5/13/2022 36880
14 5/14/2022 39393
15 5/15/2022 30418
16 5/16/2022 15220
17 5/17/2022 43684
18 5/18/2022 42053
19 5/19/2022 36577
20 5/20/2022 34825"
DF <- read.table(text = Lines)

Upvotes: 1

Related Questions