marco-thomas-dave.Rmd

---
title: "marco-thomas-dave"
output: html_document
---
# Houston Dataviz Datajam on homeprice data
By Marco, Thomas and Dave


```{r, echo=FALSE, cache=FALSE}
library(ggplot2)
library(data.table)
library(sqldf)
library(ggmap)
df <- read.csv('./Sales10k.csv',header=TRUE,sep=",",stringsAsFactors=FALSE)
sales <- data.table(df)
sales <- sales[,lon := as.numeric((GPEXT_LONGITUDE))]
sales <- sales[,lat := as.numeric((GPEXT_LATITUDE))]
sales <- sales[,lat := as.numeric((GPEXT_LATITUDE))]
sales <- sales[,CLOSEDDATE := as.Date(CLOSEDDATE, format = "%m/%d/%Y %M:%H")]
```

Histogram of home sale prices
---
```{r, echo=FALSE, cache=FALSE}
ggplot(sales) + geom_histogram(aes(x=SALESPRICE), binwidth=50000) + xlim(c(0,1e6)) + theme_classic()
```

Map - New Construction vs. Existing
---
Distrubution of New Constructions vs. Existing Constructions. This tells where
homes are being built. 
Data with out-of-range coordinates have been filtered out.
```{r, echo=FALSE, cache=FALSE}
# qmplot(lon, lat, data = downtown, maptype = "toner-background", color = I("red"))
# qmplot(lon, lat, data = sales[floor(lon) == -96][floor(lat) == 29], maptype = "toner-background", color = I("red"))
ggmap(get_map(location=c(lon=-95.75, lat=29.5))) + 
  geom_point(aes(x=lon, y=lat), data=sales[floor(lon) == -96][floor(lat) == 29]) + facet_grid(. ~ NEWCONSTRUCTION)
```


Home Sales by School District
---
This map shows in which school district homes have been sold.
```{r, echo=FALSE, cache=FALSE}
# qmplot(lon, lat, data = downtown, maptype = "toner-background", color = I("red"))
# qmplot(lon, lat, data = sales[floor(lon) == -96][floor(lat) == 29], maptype = "toner-background", color = I("red"))
ggmap(get_map(location=c(lon=-95.75, lat=29.5))) + 
  geom_point(aes(x=lon, y=lat, color=factor(SCHOOLDISTRICT)), data=sales[floor(lon) == -96][floor(lat) == 29])
```


Mean sales price in each school district
---
```{r, echo=FALSE, cache=FALSE}
# qmplot(lon, lat, data = downtown, maptype = "toner-background", color = I("red"))
# qmplot(lon, lat, data = sales[floor(lon) == -96][floor(lat) == 29], maptype = "toner-background", color = I("red"))
setkeyv(sales,c('SCHOOLDISTRICT'))
data <- sales[SALESPRICE < 1e6][floor(lon) == -96][floor(lat) == 29][,list(avg_price=mean(SALESPRICE)), by=SCHOOLDISTRICT]

ggmap(get_map(location=c(lon=-95.75, lat=29.5))) + 
  geom_point(aes(x=lon, y=lat, color=avg_price), data=sales[data]) + scale_color_continuous(low="red", high="blue") # + facet_grid(. ~ year(CLOSEDDATE))
```


Price over time by zipcode
---
What feature in the data predicts property values (over time)?

```{r, echo=FALSE, cache=FALSE}
ggplot(sales[,list(mean_price=mean(SALESPRICE)), by=list(year(CLOSEDDATE), ZIPCODE)]) + geom_line(aes(x=year, y=mean_price,color=factor(ZIPCODE),group=ZIPCODE)) + ylim(c(0, 300000))
```

Price over time by school district
---
```{r, echo=FALSE, cache=FALSE}
ggplot(sales[,list(mean_price=mean(SALESPRICE)), by=list(year(CLOSEDDATE),NEWCONSTRUCTION, SCHOOLDISTRICT)]) + geom_line(aes(x=year, y=mean_price,color=factor(SCHOOLDISTRICT),group=SCHOOLDISTRICT)) + ylim(c(0, 300000)) + facet_grid(NEWCONSTRUCTION ~ .)
```

Appendix
---
price per sqft
```{r, echo=FALSE, cache=FALSE}
ggplot(sales) + geom_point(aes(x=as.numeric(SALESPRICE) / as.numeric(SQFTBLDG), y=as.numeric(PRICESQFTSOLD) ))
```