Dealing with Dates and initial exploration

.gitignore

@@ -11,3 +11,4 @@ Data/
 # or .Rmd files
 *.pdf
 *.tex
+

Data/Crime/Crime.CPG

@@ -0,0 +1 @@
+UTF-8

Data/Crime/Crime.prj

@@ -0,0 +1 @@
+PROJCS["NAD_1983_HARN_StatePlane_Washington_North_FIPS_4601_Feet",GEOGCS["GCS_North_American_1983_HARN",DATUM["D_North_American_1983_HARN",SPHEROID["GRS_1980",6378137.0,298.257222101]],PRIMEM["Greenwich",0.0],UNIT["Degree",0.0174532925199433]],PROJECTION["Lambert_Conformal_Conic"],PARAMETER["False_Easting",1640416.666666667],PARAMETER["False_Northing",0.0],PARAMETER["Central_Meridian",-120.8333333333333],PARAMETER["Standard_Parallel_1",47.5],PARAMETER["Standard_Parallel_2",48.73333333333333],PARAMETER["Latitude_Of_Origin",47.0],UNIT["Foot_US",0.3048006096012192]],VERTCS["NAVD_1988",VDATUM["North_American_Vertical_Datum_1988"],PARAMETER["Vertical_Shift",0.0],PARAMETER["Direction",1.0],UNIT["Foot_US",0.3048006096012192]]

Data/Neighborhood/Neighborhood.cpg

@@ -0,0 +1 @@
+UTF-8

Data/Neighborhood/Neighborhood.prj

@@ -0,0 +1 @@
+PROJCS["NAD_1983_HARN_StatePlane_Washington_North_FIPS_4601_Feet",GEOGCS["GCS_North_American_1983_HARN",DATUM["D_North_American_1983_HARN",SPHEROID["GRS_1980",6378137.0,298.257222101]],PRIMEM["Greenwich",0.0],UNIT["Degree",0.0174532925199433]],PROJECTION["Lambert_Conformal_Conic"],PARAMETER["False_Easting",1640416.666666667],PARAMETER["False_Northing",0.0],PARAMETER["Central_Meridian",-120.8333333333333],PARAMETER["Standard_Parallel_1",47.5],PARAMETER["Standard_Parallel_2",48.73333333333333],PARAMETER["Latitude_Of_Origin",47.0],UNIT["Foot_US",0.3048006096012192]],VERTCS["NAVD_1988",VDATUM["North_American_Vertical_Datum_1988"],PARAMETER["Vertical_Shift",0.0],PARAMETER["Direction",1.0],UNIT["Foot_US",0.3048006096012192]]

Data/spokanepolicestations.csv

@@ -0,0 +1 @@
+Row,Lat,Lng1,47.655343,-117.4205012,47.665633,-117.4289983,47.658928,-117.4295134,47.665286,-117.4289985,47.676037,-117.4224756,47.700997,-117.4027347,47.713703,-117.4051378,47.705618,-117.3651409,47.656268,-117.37904510,47.667252,-117.43998511,47.628279,-117.369646

Station_Clustering/crime_location_by_year.Rmd

@@ -0,0 +1,42 @@
+---
+title: "Spokane Crime Location by Year"
+output: html_document
+---
+
+```{r}
+library(maps)
+library(ggmap)
+library(lubridate)
+library(dplyr)
+
+# Produce a scatterplot with crime colored by offense type for each year
+# ======================================================================
+Crime_byNeighborhood <- read.csv('Data/Spokane_CrimeData_Neighborhood.csv')   # Read in the data
+                                                                         # This will be updated
+                                                                         # when ezknitr is
+                                                                         # included
+
+#police <- read.csv('../Data/spokanepolicestations.csv')                 # Read in police station locations
+police <- data.frame(lat=c(47.665534),lng=c(-117.429014))
+
+crime <- tbl_df(data.frame(Crime_byNeighborhood)) %>%                    # Convert to local dataframe for easy printing
+  mutate(BEGINDATE = as.Date(BEGINDATE,format="%Y/%m/%d"),               # Date conversion
+         ENDDATE = as.Date(ENDDATE,format="%Y/%m/%d"),
+         year = year(BEGINDATE))                                         # Extract year
+
+lm.lng <- lm(Lng~coords.x1,data=crime %>%                                # Linear models to convert x/y to lat/long (change)
+               filter(!is.na(Lat),!is.na(Lng)))
+lm.lat <- lm(Lat~coords.x2,data=crime %>% 
+               filter(!is.na(Lat),!is.na(Lng)))
+
+crime <- crime %>% mutate(Lat=predict(lm.lat,newdata=crime),             # Use predict to update lat/long
+                 Lng=predict(lm.lng,newdata=crime))
+
+
+ggplot(county_map[county_map$subregion=="spokane",],aes(x=long, y=lat)) + geom_polygon(colour="grey",fill="grey")  + geom_point(data=crime %>% filter(year>2007) %>% sample_n(20000),aes(Lng,Lat,colour=OFFGEN)) + geom_point(data=police,aes(lng,lat),colour="black",size=2) + facet_wrap(~year)
+
+ggplot() + geom_point(data=crime %>% filter(year>2007) %>% sample_n(20000),aes(longitude,latitude,colour=OFFGEN)) + geom_point(data=police,aes(lng,lat),colour="black",size=2) + facet_wrap(~year)
+
+
+ggplot() +geom_point(data=crime %>% filter(year>2007) %>% sample_n(200),aes(coords.x1,coords.x2,colour=OFFGEN)) + facet_wrap(~year)
+```

basecode.R

@@ -1,16 +1,16 @@
-#this is an R script to import Spokane City crime data, merge it with Neighborhoods, and export it as a flat file
-rm(list=ls())
-library(rgdal)
-library(maptools)
-
-#reading in Crime shapefile
-Crime=readOGR(dsn="Crime",layer="Crime")
-#reading in Neighborhood shapefile
-Neighborhood=readOGR(dsn="Neighborhood",layer="Neighborhood")
-
-#Extracting Neighborhood designation for each point by location
-Crime_byNeighborhood=over(Crime,Neighborhood)
-#appending rest of data to each point
-Crime_byNeighborhood=spCbind(Crime,Crime_byNeighborhood)
-#saving resulting dataset as csv
-write.csv(file="Spokane_CrimeData_Neighborhood.csv",Crime_byNeighborhood)
+#this is an R script to import Spokane City crime data, merge it with Neighborhoods, and export it as a flat file
+rm(list=ls())
+library(rgdal)
+library(maptools)
+
+#reading in Crime shapefile
+Crime=readOGR(dsn="Crime",layer="Crime")
+#reading in Neighborhood shapefile
+Neighborhood=readOGR(dsn="Neighborhood",layer="Neighborhood")
+
+#Extracting Neighborhood designation for each point by location
+Crime_byNeighborhood=over(Crime,Neighborhood)
+#appending rest of data to each point
+Crime_byNeighborhood=spCbind(Crime,Crime_byNeighborhood)
+#saving resulting dataset as csv
+write.csv(file="/Data/Spokane_CrimeData_Neighborhood.csv",Crime_byNeighborhood)

spokanecrimedata.Rmd

@@ -0,0 +1,130 @@
+---
+title: "Spokane Crime - Exploring Data"
+author: "Patil"
+date: "September 26, 2016"
+output: html_document
+
+---
+
+```{r setup, include=FALSE}
+knitr::opts_chunk$set(echo = TRUE)
+```
+
+# Reading data
+
+```{r}
+# Arrive at Crime_byNeighborhood using earlier code that was originated by Krisztian
+# crimedfonly=data.frame(Crime_byNeighborhood)
+# save(crimedfonly,file="Data/Crime_byNeighborhood.Rda")
+
+load("Data/Crime_byNeighborhood.Rda")
+str(crimedfonly)
+```
+
+# Calling on a few packages
+
+```{r}
+suppressPackageStartupMessages(library(ggplot2))
+suppressPackageStartupMessages(library(dplyr))
+suppressPackageStartupMessages(library(lubridate))
+library(leaflet)
+library(ggmap)
+
+```
+
+# Total offenses
+
+```{r}
+crimedfonly %>% group_by(OFFGEN)%>%summarise(Counts=length(OFFGEN))%>% ggplot(.,aes(reorder(OFFGEN,Counts),Counts))+geom_bar(stat="identity")+geom_text(aes(x=OFFGEN,y=Counts+2000,label=Counts))+coord_flip()+theme_bw()+labs(y="Count",x="")
+
+```
+
+# Number of times offenses occured in different neighborhoods
+
+```{r}
+
+crimedfonly %>% group_by(Name)%>%summarise(Counts=length(Name))%>%filter(Name!="")%>% ggplot(.,aes(reorder(Name,Counts),Counts))+geom_bar(stat="identity")+geom_text(aes(x=Name,y=Counts+1000,label=Counts))+coord_flip()+theme_bw()+labs(y="Count",x="")
+```
+
+## Offenses in different neighborhoods
+
+```{r fig.width=10, fig.height=6}
+# A shiny app will do well here
+
+crimedfonly %>% group_by(Name,OFFGEN)%>%summarise(Counts=length(OFFGEN))%>%filter(Name!="")%>% ggplot(.,aes(OFFGEN,Counts))+geom_bar(stat="identity")+facet_wrap(~Name,scales="free")
+
+```
+
+## Messing with Dates
+
+```{r}
+# Adds 9 columns to the dataset
+
+crimedfonly=crimedfonly[,1:11]
+
+# Beginning dates parsed
+
+crimedfonly$beginyear=year(ymd(crimedfonly$BEGINDATE))
+crimedfonly$beginmonth=month(ymd(crimedfonly$BEGINDATE),label=TRUE)# label parameter inserts name of month instead of number
+crimedfonly$begindate=day(ymd(crimedfonly$BEGINDATE)) 
+crimedfonly$beginday=wday(ymd(crimedfonly$BEGINDATE)) # label parameter inserts day of week instead of number of the day in a 7-day week
+
+# Ending dates parsed
+
+crimedfonly$endyear=year(ymd(crimedfonly$ENDDATE))
+crimedfonly$endmonth=month(ymd(crimedfonly$ENDDATE),label=TRUE)
+crimedfonly$enddate=day(ymd(crimedfonly$ENDDATE))
+crimedfonly$endday=wday(ymd(crimedfonly$ENDDATE),label=TRUE)
+
+# duration, in days, between starting and ending days
+crimedfonly$durationdays= (as.duration(ymd(crimedfonly$ENDDATE)-ymd(crimedfonly$BEGINDATE)))/ddays(1)
+
+head(crimedfonly)
+
+```
+
+```{r eval=FALSE}
+crimedfonly=crimedfonly[!is.na(crimedfonly$Lat),]
+crimedfonly=crimedfonly[!is.na(crimedfonly$Lng),]
+crime2008=crimedfonly[crimedfonly$beginyear==2008,]
+
+location=c(-117.402209,47.665330)
+map=get_map(location,source="osm", color="bw",zoom=11)
+
+for (i in 1:10){
+
+
+crimedfsub=crimedfonly[crimedfonly$OFFGEN==levels(crimedfonly$OFFGEN)[i],]
+ggmap(map)+geom_point(data=crimedfsub,aes(Lng,Lat),color="red")+facet_wrap(~beginyear)+theme(panel.grid.major = element_blank(),
+                 panel.grid.minor = element_blank(),
+                axis.text = element_blank(),axis.title = element_blank(),
+                axis.ticks = element_blank())+ggtitle(paste(levels(crimedfonly$OFFGEN)[i],"Over the Years"))
+
+ggsave(file=paste("annualplot",i,".png",sep=""),scale=4)
+}
+
+```
+
+
+
+
+
+```{r eval=FALSE}
+pal <- colorFactor(rainbow(10), domain = levels(crime2008$OFFGEN))
+crimedfonly=crimedfonly[!is.na(crimedfonly$Lat),]
+crimedfonly=crimedfonly[!is.na(crimedfonly$Lng),]
+crime2008=crimedfonly[crimedfonly$beginyear==2008,]
+levels(crime2008$OFFGEN)
+leaflet(data=crime2008)%>%addTiles()%>%addCircleMarkers(color =~pal(OFFGEN),popup=~LOCATION)
+```
+
+
+
+# Few low hanging fruits to pick next
+
+
+
+* plots of trends for years, months, weeks, days, offense types
+* mapping stuff by different variables
+
+