version 1.1.0

DESCRIPTION

@@ -1,12 +1,10 @@
 Package: networktools
-Title: Assorted Tools for Identifying Important Nodes in Networks
-        (Impact, Expected Influence)
-Version: 1.0.0
+Title: Tools for Identifying Important Nodes in Networks
+Version: 1.1.0
+Date: 2017-8-8
 Authors@R: person("Payton", "Jones", email = "payton_jones@g.harvard.edu", role = c("aut", "cre"))
-Description: Includes assorted tools for network analysis. Specifically, includes functions for 
-    calculating impact statistics, which aim to identify how each node impacts 
-    the overall network structure (global strength impact, network structure impact, edge impact), 
-    and for calculating and visualizing expected influence.  
+Description: Includes assorted tools for network analysis. Useful for 
+    calculating and plotting expected influence, bridge centrality, and impact statistics. 
 Depends: R (>= 3.0.0)
 License: GPL-3
 Encoding: UTF-8
@@ -16,11 +14,11 @@ Imports:
 RoxygenNote: 6.0.1
 Suggests: knitr, rmarkdown
 VignetteBuilder: knitr
-URL: http://github.com/paytonjjones/networktools
+URL: https://CRAN.R-project.org/package=networktools
 BugReports: http://github.com/paytonjjones/networktools/issues
 NeedsCompilation: no
-Packaged: 2017-04-14 01:04:37 UTC; payto
+Packaged: 2017-08-09 16:46:12 UTC; payto
 Author: Payton Jones [aut, cre]
 Maintainer: Payton Jones <payton_jones@g.harvard.edu>
 Repository: CRAN
-Date/Publication: 2017-04-14 14:37:33 UTC
+Date/Publication: 2017-08-09 18:05:46 UTC

MD5

@@ -1,34 +1,44 @@
-d93e9ea4f3251e51aff2c3e4be04bb8b *DESCRIPTION
-055726a3bc8ef0bcc8453543471f0b5e *NAMESPACE
-bddf4091c54589c77ef555e7601956e3 *NEWS.md
+56ce9b1f5cfd79e1895907e7d82c7e38 *DESCRIPTION
+c28ad3214e9621d02eea369168de6f51 *NAMESPACE
+32c4defdfcdaf98d0ef18debd5a66cab *NEWS.md
+3d99d2c6e7b65740c7ad8a6d1d7d0ef3 *R/assumptionCheck.R
+5bd9fbab2b62362b990fe24b85807cad *R/bridge.R
 db9fae5fe89d0a82e7a600227a65791e *R/data_documentation.R
-8df274acbd65c21842c6feae69c9c292 *R/edge_impact.R
-6d61886bfbe2c4d691cee455837c6f70 *R/expectedInf.R
-2fb9555af1d5f602787fa2f4812144f6 *R/global_impact.R
-4d207de152befddd17a1ab8b1459e851 *R/hidden_functions.R
-6b8b5a97c572b3563093dcbafa691e39 *R/impact.R
-e4591f67a8ee8260f5d0302a96c13d25 *R/impact_NCT.R
-bbd3d74e355b964e4ac70407fae1b548 *R/structure_impact.R
-d4cb6f162e29b899919ded963232622b *R/summary_print_plot.R
-8a65e26936a057b6f6a39bde8315c3b2 *build/vignette.rds
+9749b1badaa667b773d82e9af8708ec5 *R/edge_impact.R
+145a2e5fd3c7c6371f91fcfa706b1435 *R/expectedInf.R
+b2885be98c23ad6c55ff538baaf3fcd8 *R/global_impact.R
+231ede57b79c0d39446d959f96737ca8 *R/hidden_functions.R
+1c85bff098685d22d00a14eb9888532e *R/impact.R
+40fe130ee122e9688390176d3b8442e9 *R/impact_NCT.R
+27ee303b044436d7f57fb0f90014ae7d *R/impact_boot.R
+f19d1aae4aca9e5cb9c6f2f960007e79 *R/networktools-package.r
+0b8b93933f1459e0d9f5d95397de60ba *R/structure_impact.R
+551bb5c7ce62d61b6c539d4535ea08b8 *R/summary_print_plot.R
+847f24ed61e7a07d1101b122ff7cbfba *build/vignette.rds
 b36c9ac17ed19012e016f719d89696a0 *data/depression.rda
 089c6b27af88285ef845229c2e78467e *data/social.rda
-3d453e9c6cb2cac3c29c76386195efea *inst/doc/Impact.R
-728388b55a75abf098bfe481d231b714 *inst/doc/Impact.Rmd
-c827e58f1d0f7fbd53affe6a21b72894 *inst/doc/Impact.pdf
-b1752dd14d507b01dadef5b00c451e7d *inst/doc/networktools.pdf
+3db79a48f7a5dd0c2ac19d5ff25d4210 *inst/doc/Impact.R
+49ac028ced27a0ea43180c810ee65886 *inst/doc/Impact.Rmd
+a5f4d37386c393ee9f8c9ab51cecc49a *inst/doc/Impact.pdf
+913a0e069d33326c12d0171058d3ca89 *inst/doc/networktools.pdf
+44951a88c1e44e76cac482e007d3abe4 *man/assumptionCheck.Rd
+411819868208601ad5e8605f0d2553c2 *man/bridge.Rd
+2f3c7f986b69779336f92dd4da52d43e *man/coerce_to_adjacency.Rd
 6b369dc2cc0ed86fa1f64b081c09b78a *man/depression.Rd
-ddeea37d0e573c342f84086a1944fda1 *man/edge.impact.Rd
+973979c2b8a3aa490f81f7ad3de76a2e *man/edge.impact.Rd
 c006aa48ab73bfabb3879dcc024cbe58 *man/expectedInf.Rd
-eb1a7a9e16423d8023c9f10cff74e635 *man/global.impact.Rd
-575ae83357ec70f5f21242f21777a201 *man/impact.NCT.Rd
-e84ebcd86e39628e59faae4fafd55e45 *man/impact.Rd
-654fe7e72cb11211412fce92006f6796 *man/plot.all.impact.Rd
-954d62892e71c27741c2dfca5b53abeb *man/plot.edge.impact.Rd
-611b36e1c85f71f37333b469157334c9 *man/plot.expectedInf.Rd
-9729b3940a2b8057f2ff99d9fee613ce *man/plot.global.impact.Rd
-c1f63b6cb912ff538a8ada0d734da89e *man/plot.structure.impact.Rd
+dfbd199423233683e1d46374d5f401c9 *man/global.impact.Rd
+2fc456793c379430e12378b37ee2407f *man/impact.NCT.Rd
+00cb8a470e756a614b7378773339d5f5 *man/impact.Rd
+e860314e7d136e76e43e6ac589b2c489 *man/impact.boot.Rd
+1156dead634fe0c301412661d9ecdfc6 *man/networktools.Rd
+622dc6d76e9c19cedbd971963dd8a130 *man/plot.all.impact.Rd
+dda646c0e050f98b4e9c802b9b6054c8 *man/plot.bridge.Rd
+16cb20ca730dcd91fdc64d7236f5e04d *man/plot.edge.impact.Rd
+29c614a613fd8dc24f48db94883d489a *man/plot.expectedInf.Rd
+d1425df1dc7d666b82e32ca2ce3c07fb *man/plot.global.impact.Rd
+8d8c76b274704cb0f52f0dce98c260a2 *man/plot.structure.impact.Rd
 1077043d1e1e5092559f7acdb02205db *man/social.Rd
-8e68c2e1fd94634f929fec7e3e2d00a8 *man/structure.impact.Rd
-dd372fe9c671fde73411860508224e1b *vignettes/Impact.Rmd
-7141c865408c67ff78ffaf4d74359e88 *vignettes/figures/pe_figure.png
+e5bea2ea7a637816fa7f527e8df55249 *man/structure.impact.Rd
+49ac028ced27a0ea43180c810ee65886 *vignettes/Impact.Rmd
+41d8ebd21578934b1c0fd6fa944dc9ad *vignettes/figures/pe_figure.png

NAMESPACE

@@ -1,23 +1,30 @@
 # Generated by roxygen2: do not edit by hand
 
 S3method(plot,all.impact)
+S3method(plot,bridge)
 S3method(plot,edge.impact)
 S3method(plot,expectedInf)
 S3method(plot,global.impact)
 S3method(plot,structure.impact)
 S3method(print,all.impact)
+S3method(print,bridge)
 S3method(print,edge.impact)
 S3method(print,expectedInf)
 S3method(print,global.impact)
 S3method(print,structure.impact)
 S3method(summary,all.impact)
+S3method(summary,bridge)
 S3method(summary,edge.impact)
 S3method(summary,expectedInf)
 S3method(summary,global.impact)
 S3method(summary,structure.impact)
+export(assumptionCheck)
+export(bridge)
+export(coerce_to_adjacency)
 export(edge.impact)
 export(expectedInf)
 export(global.impact)
 export(impact)
 export(impact.NCT)
+export(impact.boot)
 export(structure.impact)

NEWS.md

@@ -2,5 +2,8 @@
 
 * Added a `NEWS.md` file to track changes to the package.
 
+# networktools 1.1.0
 
+* The bridge() function is now available. Can be used to calculate and plot "bridge statistics", 
+  including bridge strength, bridge betweenness, bridge closeness, and bridge expected influence.
 

R/assumptionCheck.R

@@ -0,0 +1,123 @@
+#' Assumption Checking Function
+#'
+#' Checks some basic assumptions about the suitability of network analysis on
+#' your data
+#'
+#' This function is in BETA. Please report any errors.
+#'
+#' @param data dataframe or matrix of observational data (rows: observations, columns: nodes)
+#' @param type which assumptions to check? "network" tests the suitability for
+#' network analysis in general. "impact" tests the suitability for analyzing impact
+#' @param percent percent difference from grand mean that is acceptable when comparing variances.
+#' @param split if type="impact", specifies the type of split to utilize
+#' @param plot logical. Should histograms each variable be plotted?
+#' @param binary.data logical. Defaults to FALSE
+#' @param na.rm logical. Should missing values be removed?
+#'
+#' @details
+#'
+#' Network analysis rests on several assumptions. Among these:
+#' - Variance of each node is (roughly) equal
+#' - Distributions are (roughly) normal
+#'
+#' Comparing networks in impact rests on additional assumptions including:
+#' - Overall variances are (roughly) equal in each half
+#'
+#' This function checks these assumptions and notifies any violations.
+#' This function is not intended as a substitute for careful data visualization
+#' and independent assumption checks.
+#'
+#' See citations in the references section for further details.
+#'
+#' @references
+#'
+#' Terluin, B., de Boer, M. R., & de Vet, H. C. W. (2016). Differences in Connection Strength between Mental Symptoms Might Be Explained by Differences in Variance: Reanalysis of Network Data Did Not Confirm Staging. PLOS ONE, 11(11), e0155205. Retrieved from https://doi.org/10.1371/journal.pone.0155205
+#'
+#' @export
+assumptionCheck <- function(data, type=c("network", "impact"), percent=20,
+                            split=c("median","mean", "forceEqual", "cutEqual", "quartiles"),
+                            plot=FALSE, binary.data=FALSE, na.rm=TRUE) {
+if(match.arg(type)=="network"){
+  varNodes <- apply(data, 2, stats::var, na.rm=na.rm)
+  normNodes <- apply(data, 2, stats::shapiro.test)
+  normNodes <- as.numeric(sapply(normNodes,`[`,2))
+  varviolNodes <- vector()
+  normviolNodes <- vector()
+  if(is.null(colnames(data))) {colnames(data) <- 1:dim(data)[2]}
+  names(normNodes) <- names(varNodes)<-colnames(data)
+  grandMean <- mean(varNodes)
+  for(i in 1:length(varNodes)) {
+    if(abs((varNodes[i]/grandMean)-1)>(percent/100)) {varviolNodes[i] <- names(varNodes[i])}
+    if(normNodes[i]<0.01) {normviolNodes[i] <- names(varNodes[i])}
+  }
+  varviolNodes <- stats::na.omit(varviolNodes)
+  normviolNodes <- stats::na.omit(normviolNodes)
+  if(length(varviolNodes != 0)) {message("Equal node variance violation: variance of the following nodes differs more than ",
+                                percent, "% from grand mean: \n", paste0(varviolNodes, sep=" "), "\n")}
+  if(length(normviolNodes != 0)) {message("Shapiro-Wilk violation (p<0.01): Visually check normality of nodes: \n",
+                                          paste0(normviolNodes, sep=" "), "\n")}
+  cat("Grand mean of variances = ", grandMean, "\n")
+  cat("Node variances \n")
+  print(varNodes)
+  cat("\n Shapiro-Wilk Normality (p) \n")
+  print(normNodes)
+  if(plot){
+  op <- graphics::par(mfrow=c(round((dim(data)[2]/3)+.5), 3))
+  for(i in 1:dim(data)[2]) {graphics::hist(data[,i], main=NULL, xlab=colnames(data)[i], ylab=NULL)}
+  graphics::par(op)}
+}
+if(match.arg(type)=="impact"){
+  input <- stats::na.omit(data)
+  table <- as.data.frame(matrix(rep(NA, 2*dim(data)[2]), dim(data)[2], 2))
+  for(j in 1:dim(data)[2]) {
+    if(binary.data==FALSE) {
+      if(match.arg(split)=="median"){
+        hi <- input[input[,j]> stats::median(input[,j]),][,-j]
+        lo <- input[input[,j] < stats::median(input[,j]),][,-j]
+      }
+      if(match.arg(split)=="mean"){
+        hi <- input[input[,j]> mean(input[,j]),][,-j]
+        lo <- input[input[,j] < mean(input[,j]),][,-j]
+      }
+      if(match.arg(split)=="forceEqual"){
+        hi <- input[input[,j] > stats::median(input[,j]),]
+        lo <- input[input[,j] < stats::median(input[,j]),]
+        med <- input[input[,j] == stats::median(input[,j]),]
+        random_sample_med <- med[sample(1:dim(med)[1], abs(dim(hi)[1]-dim(lo)[1])),]
+        if(dim(hi)[1]<dim(lo)[1]){hi <- rbind(hi, random_sample_med)}
+        if(dim(hi)[1]>dim(lo)[1]){lo <- rbind(lo, random_sample_med)}
+        hi <- hi[,-j]
+        lo <- lo[,-j]
+      }
+      if(match.arg(split)=="cutEqual"){
+        hi <- input[input[,j] > stats::median(input[,j]),]
+        lo <- input[input[,j] < stats::median(input[,j]),]
+        if(dim(hi)[1]<dim(lo)[1]){lo <- lo[-sample(1:dim(lo)[1], abs(dim(hi)[1]-dim(lo)[1])),-j]}
+        if(dim(hi)[1]>dim(lo)[1]){hi <- hi[-sample(1:dim(hi)[1], abs(dim(hi)[1]-dim(lo)[1])),-j]}
+      }
+      if(match.arg(split)=="quartile"){
+        hi <- input[input[,j]>= stats::quantile(input[,j],probs=.75),][,-j]
+        lo <- input[input[,j]<= stats::quantile(input[,j],probs=.25),][,-j]
+      }
+    }
+    if(binary.data==TRUE)  {
+      hi <- input[input[,j]==1,][,-j]
+      lo <- input[input[,j]==0,][,-j]
+      if(match.arg(split)=="cutEqual"){
+        if(dim(hi)[1]<dim(lo)[1]){lo <- lo[-sample(1:dim(lo)[1], abs(dim(hi)[1]-dim(lo)[1])),]}
+        if(dim(hi)[1]>dim(lo)[1]){hi <- hi[-sample(1:dim(hi)[1], abs(dim(hi)[1]-dim(lo)[1])),]}
+      }
+    }
+    table[j,1] <- sum(diag(stats::var(lo)))
+    table[j,2] <- sum(diag(stats::var(hi)))
+    if(abs((table[j,1]/table[j,2])-1)>(percent/100)) {message("Equal network variances violation: difference in variance between halves on node \"",
+                                                         colnames(data)[j], "\" is greater than ", percent)}
+  }
+  colnames(table) <- c("Low Half (variance)", "High Half (variance)")
+  rownames(table) <- colnames(data)
+  cat("Total network variances \n")
+  table
+}
+}
+
+