PopSyn3vsTom.Rmd

```{r echo=FALSE, message=FALSE, warning=FALSE}
library(dplyr)
library(ggplot2)
library(reshape2)
library(sp)
library(rgdal)

#############################################################################################
```


```{r init, echo=FALSE, message=FALSE, warning=FALSE}
#############################################################################################
# set working directory

wd <- setwd("c:/personal/r")
```

####  This R Markdown file does a comparison of Tom's macroeconomic model estimates for 2011 with those prepared by the MTO and used in PopSYn3. 

##### Read all the files necessary, including shapefiles for getting equivalency tables that map TAZs in the GGHM to CSDs

```{r }
csd <- readOGR(wd, "CSD")
  csd.df <- csd@data
taz <- readOGR(wd, "GGHM_Cen")
  taz.df <- taz@data %>% subset(., select = c("TAZ_NO", "CSDUID"))

pop3 <- read.csv(file = "synpop_person_with_required_fields.csv", stringsAsFactors = FALSE)
hh3 <- read.csv(file = "synpop_hh.csv", stringsAsFactors = FALSE)

# read Tom's numbers
tom <- read.csv(file = "TomNumbers.csv", stringsAsFactors = FALSE)
  tom <- merge(tom, csd.df, by.x = "CSDUID", by.y = "CSDUID", all.x = T) %>% 
    subset(., select = c("CSDUID", "TomPop", "TomHH", "GGH"))
```

Summarise and count PopSyn3 population data

```{r summarize Popsyn3 Population data}
#' create summaries
pop3.sum <- pop3 %>% group_by(taz) %>% summarise(popl3 = sum(finalweight))
  
  #' join to bring in taz level mapping
  pop3.sum <- merge(pop3.sum, taz.df, by.x = "taz", by.y = "TAZ_NO", all.x = T) %>% 
    transform(., CSDUID1 = as.numeric(as.character(CSDUID))) %>% 
    subset(., select = c("taz", "popl3", "CSDUID1"))
    
  #' get rid of NAs and populate the entire CSD column
    pop3.sum[is.na(pop3.sum)] <- 0
      pop3.sum <- transform(pop3.sum, CSDUID1 = ifelse(CSDUID1 == 0, taz, CSDUID1)) 
        pop3.sum <- pop3.sum %>% group_by(CSDUID1) %>% summarise(popl3 = sum(popl3))
    
        
```


Summarize and count Popsyn3 Household data

```{r summarize Popsyn3 Household data}
#' create summaries
hh3.sum <- hh3 %>% group_by(taz) %>%
  summarise(hh3 = sum(finalweight))
 
  #' join to bring in taz level mapping
  hh3.sum <- merge(hh3.sum, taz.df, by.x = "taz", by.y = "TAZ_NO", all.x = T) %>% 
    transform(., CSDUID1 = as.numeric(as.character(CSDUID))) %>% 
    subset(., select = c("taz", "hh3", "CSDUID1"))
    
  #' get rid of NAs and populate the entire CSD column
    hh3.sum[is.na(hh3.sum)] <- 0
      hh3.sum <- transform(hh3.sum, CSDUID1 = ifelse(CSDUID1 == 0, taz, CSDUID1)) 
        hh3.sum <- hh3.sum %>% group_by(CSDUID1) %>% summarise(hh3 = sum(hh3))
```

Now join the datasets together and estimate their differences

```{r Join the datasets and estimate differences, echo=FALSE, message=FALSE, warning=FALSE}
joined1 <- merge(tom, pop3.sum, by.x="CSDUID", by.y="CSDUID1", all.x = T) 

#now join the housing data
joined1 <- merge(joined1, hh3.sum, by.x="CSDUID", by.y="CSDUID1", all.x = T)
  joined1 <- transform(joined1, popdiff = TomPop - popl3) %>% transform(., hhdiff = TomHH - hh3)
  joined1[is.na(joined1)] <- 0
  
  print(paste0("The difference between Tom's population estimates and those from PopSyn3 are ", sum(joined1$popdiff)))
  print(paste0("The difference between Tom's household estimates and those from PopSyn3 are ", sum(joined1$hhdiff))) 
  
 joined1.ggh <- subset(joined1, GGH == 1)
  
  print(paste0("The difference between Tom's population estimates and those from PopSyn3 within the GGH are ", sum(joined1.ggh$popdiff)))
  print(paste0("The difference between Tom's household estimates and those from PopSyn3 within the GGH are ", sum(joined1.ggh$hhdiff)))

```


```{r Plot differences between input and output fields}

# plot the differences of households. GGH = 1 and Outside GGH = 0
ggplot(joined1, aes(x=CSDUID, y=hhdiff))+geom_line(color="grey")+geom_point(color="red")+
  ggtitle("Household Differences (Tom Households - PopSyn3 Households)") + facet_grid(. ~ GGH)

# plot the differences of population.  GGH = 1 and Outside GGH = 0
ggplot(joined1, aes(x=CSDUID, y=popdiff))+geom_line(color="grey")+geom_point(color="red")+
  ggtitle("Population Differences (Tom Population - PopSyn3 Population)") + facet_grid(. ~ GGH)
```
```