1 files changed, 48 insertions, 8 deletions
diff --git a/R Scripts/data-prep.R b/R Scripts/data-prep.R
index 127acca..ca2fdc2 100755
--- a/R Scripts/data-prep.R
+++ b/R Scripts/data-prep.R
@@ -80,6 +80,18 @@ for (i in 1:dim(a)[1]){
   }
 }
 
+# clean up entries where sex is missing
+a = arrests[arrests$sex_code_cd=='X',]
+for (i in 1:dim(a)[1]){
+  ir = a$ir_no[i]
+  arr = arrests[arrests$ir_no==ir,]
+  arr = arr[arr$sex_code_cd != 'X',]
+  if(dim(arr)[1]>0){
+    arrests$sex_code_cd[as.numeric(rownames(a[i,]))] = names(which.max(table(arr$sex_code_cd)))
+  }
+} 
+arrests$sex_code_cd[arrests$sex_code_cd=='X'] = 'M'
+
 #I need to add the "ir" for this to make sense when I "project"
 arrests$ir2 <- paste("ir", arrests$ir_no)
 
@@ -171,6 +183,17 @@ murders = murders[match(unique(murders$VICTIM_IR_NO),murders$VICTIM_IR_NO),]
 murders = murders[as.Date(murders$INJURY_DATE,format='%m/%d/%y')>=start_date,]
 murders$ir2 = paste("ir", murders$VICTIM_IR_NO)
 
+# clear nonfatals that led to death
+v = victims[victims$IR_NO %in% murders$VICTIM_IR_NO,]
+rows = c()
+for(i in 1:dim(v)[1]){
+  row = which(rownames(victims)==as.numeric(rownames(v[i,])))
+  m = murders[murders$VICTIM_IR_NO==v$IR_NO[i],]
+  dup = as.Date(v$INCIDENT_DATE[i],format='%m/%d/%y') %in% as.Date(m$INJURY_DATE,format='%m/%d/%y')
+  if(dup==T) rows = c(rows,row)
+}
+victims = victims[-rows,]
+
 # set victim data in network
 vtab = as.data.frame(table(victims$ir2))
 match_vector = match(V(person)$name,vtab$Var1)
@@ -193,7 +216,8 @@ for(i in 1:length(vics)){
   if (i%%3000==0) print(i)
   name = vics[i]
   ids = which(match_vector==name)
-  dates = sort(as.Date(victims$INCIDENT_DATE[ids],format='%m/%d/%y'))
+  dates = unique(sort(as.Date(victims$INCIDENT_DATE[ids],format='%m/%d/%y')))
+#   if(!is.na(V(person)$fatal_date[i])) dates = dates[dates != V(person)$fatal_date[ids]]
   nfd1[i] = as.character(dates[1])
   nfd2[i] = as.character(dates[2])
   nfd3[i] = as.character(dates[3])
@@ -211,6 +235,7 @@ V(person)$nonfatal_date_3[vics] = nfd3
 V(person)$nonfatal_date_4[vics] = nfd4
 V(person)$nonfatal_date_5[vics] = nfd5
 
+
 # convert dates into numeric values ("days")
 start_date
 V(person)$fatal_day = as.numeric(as.Date(V(person)$fatal_date)-start_date)
@@ -223,25 +248,40 @@ V(person)$nonfatal_day_5 = as.numeric(as.Date(V(person)$nonfatal_date_5)-start_d
 #===================================================================
 # set gang attributes
 gangs <- read.csv("Sept2014-ganglist.csv", header=T)
-gangs = gangs[match(unique(gangs$IR_NO),gangs$IR_NO),]
 gangs$ir2 <- paste("ir", gangs$IR_NO)
 
+t = table(gangs$IR_NO)
+t = t[t>1]
+irs = as.numeric(attr(t,'name'))
+for(ir in irs){
+  if(which(ir==irs)%%1000==0)print(which(ir==irs))
+  g = gangs[gangs$IR_NO==ir,]
+  gangs$GANG_NAME[as.numeric(rownames(g))] = names(which.max(table(g$GANG_NAME)))
+}
+
+gangs = gangs[match(unique(gangs$IR_NO),gangs$IR_NO),]
+gnames = as.character(gangs$GANG_NAME)
+gnames[is.na(gnames)] = 'Unknown'
+
 V(person)$gang.member <- V(person)$name %in% gangs$ir2
 
 match_vector = match(V(person)$name, gangs$ir2)
-gnames = gangs$GANG_NAME[match_vector]
-gnames = as.character(gnames)
-gnames[V(person)$gang.member==''] = 'Unknown'
+gnames = gnames[match_vector]
 gnames[V(person)$gang.member==F] = 'None'
 V(person)$gang.name <- as.character(gnames)
-V(person)$faction.name <- as.character(gangs$FACTION_NAME[match_vector])
+# V(person)$faction.name <- as.character(gangs$FACTION_NAME[match_vector])
+
+# clean up later to make this fit with process
+t = table(V(person)$gang.name)
+gs = names(t)[t<50]
+V(person)$gang.name[V(person)$gang.name %in% gs] = 'Unknown'
 
 #===================================================================
 
 # save data
 person = remove.edge.attribute(person,'weight')
 # person_data = get.data.frame(person,'both')
-save(person, file="chi-9sep2015.RData")
+save(person, file="chi-14sep2015.RData")
 
 #===================================================================
 # get LCC of the network
@@ -254,7 +294,7 @@ lcc_edges = as_data_frame(lcc,'edges')
 
 # update lcc_verts
 lcc_verts = get.data.frame(lcc,'vertices')
-lcc_verts = lcc_verts[,c(1,23,24,2:22)]
+lcc_verts = lcc_verts[,c(1,23,24,2:21)]
 
 # save file
 save(lcc, lcc_verts, lcc_edges, vic_ids, file="lcc.RData")