SpatialAnalysisCode.Rmd

---
title: "Spatial Taxi Code"
author: "Emilie Bruzelius"
date: "April 22, 2015"
output: html_document
---

```{r}
rm(list=ls())
```

Read and plot block shapefile
```{r}
library(maptools)
library(ggplot2)

block <- readShapePoly('/Users/emiliebruzelius/Desktop/git/taxi/taxi/nycb2010_15a/nycb2010.shp')

ggplot() +  geom_polygon(data=block, 
                         aes(x=long, y=lat, group=group))
```

Create neighbors
```{r}
library(sp)
library(spdep)
length(block)
block.nb <- poly2nb(block)
head(block.nb)
plot(block.nb, coordinates(block), 
     col = "red", pch = ".")
plot(block, add = T)

coords <- coordinates(block)
IDs <- row.names(coords)

block.nb3 <- graph2nb(gabrielneigh(coords), 
                      row.names = IDs)
plot(block.nb3, coords, pch = ".", col = "red")
plot(block.nb3, add = T)
```

Create block weights
```{r}
block.wts <- nb2listw(block.nb3, 
                      zero.policy = T, style = "B")
print(block.wts, zero.policy = T)
names(block.wts)
```

Moran's Scatterplot
```{r}
set.seed(987654)
n <- length(block.nb3)
uncorr.x <- rnorm(n)
rho <- 0.05
autocorr.x <- invIrW(block.wts, rho, 
                     feasible = TRUE) %*% uncorr.x
plot(autocorr.x, lag(nycZIPS.wts, autocorr.x), 
     xlab = "autocorrelated random variable",
     ylab = "spatial lag", 
     main = "Autocorrelated random variable",
     cex.main = 0.8, cex.lab = 0.8)
```

moran.plot(outcome.variable, listw = nb2listw(neighbor.list))

Moran's I
```{r}
moran.x <- moran.test(autocorr.x, 
                      listw = block.wts, zero.policy = T)
moran.x
summary(moran.x)
moran.x$p.value
```

Code for geoid
```{r}
https://github.com/nygeog/nyc_gis_tools/blob/master/calc_geoid_from_boro.py\
```

Merge
```{r}
library(sp)
block.df <- merge(blocks, taxi, all.x = T, all.y = F, by.x = "geoid", by.y = "geoid")
```

```{r}
###moran.cases <- moran.test(nycZIPS2$cases, 
                          #listw = block.wts, zero.policy = TRUE)
#moran.cases
```