## Free Statistics

of Irreproducible Research!

Author's title
Author*Unverified author*
R Software Modulerwasp_meanplot.wasp
Title produced by softwareMean Plot
Date of computationMon, 07 Jan 2008 06:38:24 -0700
Cite this page as followsStatistical Computations at FreeStatistics.org, Office for Research Development and Education, URL https://freestatistics.org/blog/index.php?v=date/2008/Jan/07/t119971350673encrjvyxgcqw7.htm/, Retrieved Fri, 24 Mar 2023 13:12:35 +0000
Statistical Computations at FreeStatistics.org, Office for Research Development and Education, URL https://freestatistics.org/blog/index.php?pk=7917, Retrieved Fri, 24 Mar 2023 13:12:35 +0000
QR Codes:

Original text written by user:
IsPrivate?No (this computation is public)
User-defined keywordsmean, median, midrange, mean plot, seasonality, trend
Estimated Impact1203
Family? (F = Feedback message, R = changed R code, M = changed R Module, P = changed Parameters, D = changed Data)
-     [Bivariate Data Series] [Bivariate dataset] [2008-01-05 23:51:08] [74be16979710d4c4e7c6647856088456]
F RMPD    [Mean Plot] [Colombia Coffee] [2008-01-07 13:38:24] [d41d8cd98f00b204e9800998ecf8427e] [Current]
- RMPD      [Harrell-Davis Quantiles] [steve] [2008-04-15 18:01:58] [74be16979710d4c4e7c6647856088456]
-   PD      [Mean Plot] [Notched boxplot -...] [2008-12-07 09:49:05] [c45c87b96bbf32ffc2144fc37d767b2e]
- RMPD        [Central Tendency] [central tendency] [2008-12-14 12:46:40] [c45c87b96bbf32ffc2144fc37d767b2e]
-    D          [Central Tendency] [trimmed mean] [2008-12-16 21:19:25] [c45c87b96bbf32ffc2144fc37d767b2e]
-    D        [Mean Plot] [Notched boxplot -...] [2008-12-14 12:56:52] [74be16979710d4c4e7c6647856088456]
-    D          [Mean Plot] [Notched boxplot -...] [2008-12-16 21:22:54] [c45c87b96bbf32ffc2144fc37d767b2e]
-  M D            [Mean Plot] [] [2009-12-30 10:19:42] [d2d412c7f4d35ffbf5ee5ee89db327d4]
-    D              [Mean Plot] [] [2009-12-30 14:08:44] [d2d412c7f4d35ffbf5ee5ee89db327d4]
-   PD      [Mean Plot] [Mean Plot - Xt - ...] [2008-12-22 10:24:52] [33f4701c7363e8b81858dafbf0350eed]
-    D        [Mean Plot] [Mean Plot - Trans...] [2008-12-22 13:53:01] [33f4701c7363e8b81858dafbf0350eed]
-    D          [Mean Plot] [Mean Plot - Trans...] [2008-12-22 20:19:36] [b187fac1a1b0cb3920f54366df47fea3]
-    D          [Mean Plot] [mean plot - trans...] [2008-12-22 20:40:06] [b641c14ac36cb6fee377f3b099dcac19]
-    D        [Mean Plot] [Mean Plot - Xt - ...] [2008-12-22 20:11:47] [b187fac1a1b0cb3920f54366df47fea3]
-    D        [Mean Plot] [mean plot - Xt - ...] [2008-12-22 20:23:09] [b641c14ac36cb6fee377f3b099dcac19]
-  MPD      [Mean Plot] [Workshop 6 assign...] [2010-11-05 08:31:09] [56d90b683fcd93137645f9226b43c62b]
-  MPD      [Mean Plot] [W6 Assignments Tu...] [2010-11-05 08:36:37] [56d90b683fcd93137645f9226b43c62b]
-  MPD      [Mean Plot] [WS6: Toturial ass...] [2010-11-05 10:06:39] [1fd136673b2a4fecb5c545b9b4a05d64]
-   P         [Mean Plot] [ws6.2 tutorial (s...] [2010-11-15 08:09:29] [e4076051fbfb461c886b1e223cd7862f]
-   P         [Mean Plot] [ws6.2 tutorial (f...] [2010-11-15 08:14:30] [e4076051fbfb461c886b1e223cd7862f]
- RM          [Mean Plot] [seasonality] [2011-11-15 14:45:19] [d31984dff2665bea309b726bae3d5241]
- RM          [Mean Plot] [forecasts acceptable] [2011-11-15 14:48:59] [d31984dff2665bea309b726bae3d5241]
- RM          [Mean Plot] [] [2011-11-15 21:53:28] [74be16979710d4c4e7c6647856088456]
- RM          [Mean Plot] [] [2011-11-15 21:59:12] [74be16979710d4c4e7c6647856088456]
-  MPD      [Mean Plot] [Retailprijs - Sei...] [2010-11-05 10:26:46] [aeb27d5c05332f2e597ad139ee63fbe4]
-    D        [Mean Plot] [Mean Plot - Niet ...] [2010-11-12 11:43:04] [aeb27d5c05332f2e597ad139ee63fbe4]
-    D          [Mean Plot] [Mean Plot – Nie...] [2010-12-17 13:44:39] [aeb27d5c05332f2e597ad139ee63fbe4]
-    D        [Mean Plot] [Mean Plot - VDAB ...] [2010-11-12 11:44:22] [aeb27d5c05332f2e597ad139ee63fbe4]
- RMPD      [Mean Plot] [WS6 - Simple Line...] [2010-11-05 13:17:47] [1f5baf2b24e732d76900bb8178fc04e7]
- RMPD      [Mean Plot] [Yt Seasonality] [2010-11-05 16:54:03] [97ad38b1c3b35a5feca8b85f7bc7b3ff]
-   PD        [Mean Plot] [Vraag 2: seasonality] [2010-11-11 09:55:50] [39c51da0be01189e8a44eb69e891b7a1]
- R PD        [Mean Plot] [] [2010-11-16 17:11:22] [8ef75e99f9f5061c72c54640f2f1c3e7]
F    D        [Mean Plot] [ws6] [2010-11-17 09:01:38] [f9eaed74daea918f73b9f505c5b1f19e]
F    D        [Mean Plot] [ws6] [2010-11-17 09:03:07] [f9eaed74daea918f73b9f505c5b1f19e]
- RM          [Mean Plot] [WS 6 - 9] [2011-11-15 14:49:53] [74be16979710d4c4e7c6647856088456]
- R  D        [Mean Plot] [C6 A2.2] [2011-11-15 15:38:45] [d1ce18d003fa52f731d1c3ce8b58d5f9]
- R PD        [Mean Plot] [] [2011-11-15 20:18:15] [ec2187f7727da5d5d939740b21b8b68a]
- R PD        [Mean Plot] [] [2011-11-15 20:19:06] [ec2187f7727da5d5d939740b21b8b68a]
-  MPD      [Mean Plot] [WS6 - Assignment ...] [2010-11-06 11:08:15] [8ef49741e164ec6343c90c7935194465]
-   P         [Mean Plot] [WS6 assignment 2 ...] [2010-11-16 20:16:02] [8214fe6d084e5ad7598b249a26cc9f06]
-   P         [Mean Plot] [WS6 assignment 2 ...] [2010-11-16 20:17:31] [8214fe6d084e5ad7598b249a26cc9f06]
-   P           [Mean Plot] [] [2010-11-17 09:46:18] [afdb2fc47981b6a655b732edc8065db9]
-  MPD      [Mean Plot] [] [2010-11-06 13:07:35] [39e83c7b0ac936e906a817a1bb402750]
- RMPD      [Mean Plot] [Mean Plot Car sales] [2010-11-06 16:22:27] [97ad38b1c3b35a5feca8b85f7bc7b3ff]
-    D        [Mean Plot] [Mean Plot Gasolin...] [2010-11-06 16:47:42] [97ad38b1c3b35a5feca8b85f7bc7b3ff]
- R  D          [Mean Plot] [] [2011-11-15 17:01:45] [06c08141d7d783218a8164fd2ea166f2]
- R P           [Mean Plot] [] [2011-11-15 21:34:25] [ec2187f7727da5d5d939740b21b8b68a]
- R  D        [Mean Plot] [] [2011-11-15 15:57:49] [06c08141d7d783218a8164fd2ea166f2]
-    D          [Mean Plot] [] [2011-12-20 18:10:55] [06c08141d7d783218a8164fd2ea166f2]

[Truncated]
Feedback Forum

Post a new message
Dataseries X:
87.28
87.28
87.09
86.92
87.59
90.72
90.69
90.3
89.55
88.94
88.41
87.82
87.07
86.82
86.4
86.02
85.66
85.32
85
84.67
83.94
82.83
81.95
81.19
80.48
78.86
69.47
68.77
70.06
73.95
75.8
77.79
81.57
83.07
84.34
85.1
85.25
84.26
83.63
86.44
85.3
84.1
83.36
82.48
81.58
80.47
79.34
82.13
81.69
80.7
79.88
79.16
78.38
77.42
76.47
75.46
74.48
78.27
80.7
79.91
78.75
77.78
81.14
81.08
80.03
78.91
78.01
76.9
75.97
81.93
80.27
78.67
77.42
76.16
74.7
76.39
76.04
74.65
73.29
71.79
74.39
74.91
74.54
73.08
72.75
71.32
70.38
70.35
70.01
69.36
67.77
69.26
69.8
68.38
67.62
68.39
66.95
65.21
66.64
63.45
60.66
62.34
60.32
58.64
60.46
58.59
61.87
61.85
67.44
77.06
91.74
93.15
94.15
93.11
91.51
89.96
88.16
86.98
88.03
86.24
84.65
83.23
81.7
80.25
78.8
77.51
76.2
75.04
74
75.49
77.14
76.15
76.27
78.19
76.49
77.31
76.65
74.99
73.51
72.07
70.59
71.96
76.29
74.86
74.93
71.9
71.01
77.47
75.78
76.6
76.07
74.57
73.02
72.65
73.16
71.53
69.78
67.98
69.96
72.16
70.47
68.86
67.37
65.87
72.16
71.34
69.93
68.44
67.16
66.01
67.25
70.91
69.75
68.59
67.48
66.31
64.81
66.58
65.97
64.7
64.7
60.94
59.08
58.42
57.77
57.11
53.31
49.96
49.4
48.84
48.3
47.74
47.24
46.76
46.29
48.9
49.23
48.53
48.03
54.34
53.79
53.24
52.96
52.17
51.7
58.55
78.2
77.03
76.19
77.15
75.87
95.47
109.67
112.28
112.01
107.93
105.96
105.06
102.98
102.2
105.23
101.85
99.89
96.23
94.76
91.51
91.63
91.54
85.23
87.83
87.38
84.44
85.19
84.03
86.73
102.52
104.45
106.98
107.02
99.26
94.45
113.44
157.33
147.38
171.89
171.95
132.71
126.02
121.18
115.45
110.48
117.85
117.63
124.65
109.59
111.27
99.78
98.21
99.2
97.97
89.55
87.91
93.34
94.42
93.2
90.29
91.46
89.98
88.35
88.41
82.44
79.89
75.69
75.66
84.5
96.73
87.48
82.39
83.48
79.31
78.16
72.77
72.45
68.46
67.62
68.76
70.07
68.55
65.3
58.96
59.17
62.37
66.28
55.62
55.23
55.85
56.75
50.89
53.88
52.95
55.08
53.61
58.78
61.85
55.91
53.32
46.41
44.57
50
50
53.36
46.23
50.45
49.07
45.85
48.45
49.96
46.53
50.51
47.58
48.05
46.84
47.67
49.16
55.54
55.82
58.22
56.19
57.77
63.19
54.76
55.74
62.54
61.39
69.6
79.23
80
93.68
107.63
100.18
97.3
90.45
80.64
80.58
75.82
85.59
89.35
89.42
104.73
95.32
89.27
90.44
86.97
79.98
81.22
87.35
83.64
82.22
94.4
102.18


 Summary of compuational transaction Raw Input view raw input (R code) Raw Output view raw output of R engine Computing time 4 seconds R Server 'George Udny Yule' @ 72.249.76.132

\begin{tabular}{lllllllll}
\hline
Summary of compuational transaction \tabularnewline
Raw Input & view raw input (R code)  \tabularnewline
Raw Output & view raw output of R engine  \tabularnewline
Computing time & 4 seconds \tabularnewline
R Server & 'George Udny Yule' @ 72.249.76.132 \tabularnewline
\hline
\end{tabular}
%Source: https://freestatistics.org/blog/index.php?pk=7917&T=0

[TABLE]
[ROW][C]Summary of compuational transaction[/C][/ROW]
[ROW][C]Raw Input[/C][C]view raw input (R code) [/C][/ROW]
[ROW][C]Raw Output[/C][C]view raw output of R engine [/C][/ROW]
[ROW][C]Computing time[/C][C]4 seconds[/C][/ROW]
[ROW][C]R Server[/C][C]'George Udny Yule' @ 72.249.76.132[/C][/ROW]
[/TABLE]
Source: https://freestatistics.org/blog/index.php?pk=7917&T=0

Globally Unique Identifier (entire table): ba.freestatistics.org/blog/index.php?pk=7917&T=0

As an alternative you can also use a QR Code:

The GUIDs for individual cells are displayed in the table below:

 Summary of compuational transaction Raw Input view raw input (R code) Raw Output view raw output of R engine Computing time 4 seconds R Server 'George Udny Yule' @ 72.249.76.132

par1 <- as.numeric(par1)(n <- length(x))(np <- floor(n / par1))arr <- array(NA,dim=c(par1,np+1))darr <- array(NA,dim=c(par1,np+1))ari <- array(0,dim=par1)dx <- diff(x)j <- 0for (i in 1:n){j = j + 1ari[j] = ari[j] + 1arr[j,ari[j]] <- x[i]darr[j,ari[j]] <- dx[i]if (j == par1) j = 0}ariarrdarrarr.mean <- array(NA,dim=par1)arr.median <- array(NA,dim=par1)arr.midrange <- array(NA,dim=par1)for (j in 1:par1){arr.mean[j] <- mean(arr[j,],na.rm=TRUE)arr.median[j] <- median(arr[j,],na.rm=TRUE)arr.midrange[j] <- (quantile(arr[j,],0.75,na.rm=TRUE) + quantile(arr[j,],0.25,na.rm=TRUE)) / 2}overall.mean <- mean(x)overall.median <- median(x)overall.midrange <- (quantile(x,0.75) + quantile(x,0.25)) / 2bitmap(file='plot1.png')plot(arr.mean,type='b',ylab='mean',main='Mean Plot',xlab='Periodic Index')mtext(paste('#blocks = ',np))abline(overall.mean,0)dev.off()bitmap(file='plot2.png')plot(arr.median,type='b',ylab='median',main='Median Plot',xlab='Periodic Index')mtext(paste('#blocks = ',np))abline(overall.median,0)dev.off()bitmap(file='plot3.png')plot(arr.midrange,type='b',ylab='midrange',main='Midrange Plot',xlab='Periodic Index')mtext(paste('#blocks = ',np))abline(overall.midrange,0)dev.off()bitmap(file='plot4.png')z <- data.frame(t(arr))names(z) <- c(1:par1)(boxplot(z,notch=TRUE,col='grey',xlab='Periodic Index',ylab='Value',main='Notched Box Plots - Periodic Subseries'))dev.off()bitmap(file='plot4b.png')z <- data.frame(t(darr))names(z) <- c(1:par1)(boxplot(z,notch=TRUE,col='grey',xlab='Periodic Index',ylab='Value',main='Notched Box Plots - Differenced Periodic Subseries'))dev.off()bitmap(file='plot5.png')z <- data.frame(arr)names(z) <- c(1:np)(boxplot(z,notch=TRUE,col='grey',xlab='Block Index',ylab='Value',main='Notched Box Plots - Sequential Blocks'))dev.off()bitmap(file='plot6.png')z <- data.frame(cbind(arr.mean,arr.median,arr.midrange))names(z) <- list('mean','median','midrange')(boxplot(z,notch=TRUE,col='grey',ylab='Overall Central Tendency',main='Notched Box Plots'))dev.off()