index.qmd

---
title: "Trends in home and lot size over time in Chapel Hill"
format: 
  html:
    title-block-banner: true
    toc: true
    toc-location: left
    toc-depth: 3
    html-math-method: katex
    code-fold: true
    code-summary: "show the code"
    code-overflow: wrap
    code-tools: true
    number-sections: true
    theme: 
      dark: slate
      light: flatly
    fig-width: 9
    fig-height: 6
editor: visual
---

## Introduction

Let's look at the trend in the lot size and home size of houses built in Chapel Hill over time.

## Get some data

We need: Year Built Structure Type, Parcel Size, and Floor area.

Let's pull in the Orange County parcel data, which tells us where parcels are and their area. It's available [here](ParcelsGDB.zip).

We also need the Orange County Address data, which has the year built, floor area. That's available [here](%22https://web.co.orange.nc.us/gisdownloads/addresses.zip%22).

Finally, we need the municipal boundaries to accurately capture which jurisdictions are relevant. That's available [here](%22https://web.co.orange.nc.us/gisdownloads/city.zip%22).

```{r, output=FALSE, echo=FALSE, message=FALSE, warning=FALSE}
library(tidyverse)
library(Hmisc)
library(mdbr)
library(knitr)
library(sf)
library(gganimate)
```

```{r, eval = FALSE}
download.file("https://web.co.orange.nc.us/gisdownloads/ParcelsGDB.zip", destfile = "ParcelsGDB.zip")
unzip("ParcelsGDB.zip")

download.file("https://web.co.orange.nc.us/gisdownloads/addresses.zip", destfile = "addresses.zip")
unzip("addresses.zip")

download.file("https://web.co.orange.nc.us/gisdownloads/city.zip", destfile = "city.zip")
unzip("city.zip")
```

## Read in the data!

And get only the Chapel Hill boundary.

```{r}
data <- read_mdb("ParcelsGDB.mdb", table="PARCELVIEW")
addr <- read_sf("addresses.shp")
city <- read_sf("city.shp")
ch <- city %>% filter(CITYCODE == "CH") %>% st_union() %>% st_as_sf()
```

## Summarise the data

First, we filter the addresses to only single family-type residences.

```{r, results="asis"}
addr <- addr %>% filter(LBCSSubCod < 1160 & LBCSSubCod >0) 
addr <- addr[ch,]
#mapview::mapview(addr,zcol="LBCSDesc")
parcels <- sf::read_sf("All_Parcels.geojson")

duplex_footprint <- 2774

d <- right_join(parcels,st_drop_geometry(addr),by="PIN") %>% 
  filter(LBCSDesc %in% c("Res-Single Family", "Res-MH-SW", "Res-MH-DW-TW")) %>% 
  filter(OwnerAddress!="PROPERTY OFFICE UNC") %>% 
  st_transform(2264) %>% st_make_valid() %>% 
  mutate(area_acr = as.numeric(st_area(.)/43560), area_ft2 = as.numeric(st_area(.))) %>% 
  filter(area_acr<5) %>% 
  mutate(plexable = case_when(area_ft2>=4*duplex_footprint ~ "yes", TRUE ~ "no"))

```

We then filter to only parcels in CH, and animate the plot.

```{r, results="asis"}
data2 <- data %>% filter(PIN %in% d$PIN)
gif <- data2 %>% 
filter(FinArea < 6000 & YearBlt > 1900 & Shape_Area < 43560*2) %>% 
  arrange(desc(YearBlt)) %>%
  ggplot(aes(x=Shape_Area/43560, y=FinArea, color=YearBlt)) +
    geom_point(alpha=0.5) + 
    transition_time(YearBlt) + 
    labs(title = 'Year: {frame_time}') + 
    shadow_mark() + 
    scale_colour_viridis_b() + 
    xlab("Lot Size (Acres)") + 
    ylab("Unit sqft")

gif<-animate(gif, renderer=gifski_renderer())

anim_save("gig.gif",gif)
gif
```