---
title: "Charlotte Beer Tweets - Leaflet Demo"
output: 
  flexdashboard::flex_dashboard:
    storyboard: true
    social: menu
    source: embed
---

```{r setup, include=FALSE}
library(flexdashboard); library(leaflet); library(htmltools); library(KernSmooth); library(sp)

tweets <- read.csv("./CLT_beer_tweets.csv", stringsAsFactors = F)
#tweets <- read.csv("./CharlotteTweets20Sample.csv", stringsAsFactors = F)

points <- subset(tweets, !is.na(point_long))
```

### Leaflet Demo: Simple plot with only points.

```{r}
leaflet(points) %>%
  addTiles() %>%
  addCircleMarkers(lng=points$point_long, lat=points$point_lat)

```

***

Question: Where's the best area to get a beer in Charlotte? 

- Let's use as a proxy the number of mentions by geotags on Twitter.

- We'll use [Leaflet](https://rstudio.github.io/leaflet/) to map the tweets.

- It's important to know that there are two types of geolocation tags: point and polygon (place).

- Points are exact lat/long; polygons are lat/long bounding boxes for default places (e.g. Charlotte, NC).

- In this exercise, we'll only look at the point Tweets.

- The data is from Dec 2015 - Feb 2016.


### Point enhancements: Adjust size, opacity and add in pop-ups of tweet text

```{r}
leaflet(points) %>%
  addTiles() %>%
  addCircleMarkers(lng=points$point_long, lat=points$point_lat,   popup = points$body, #color = pal(points$generator),
             stroke = FALSE, fillOpacity = 0.5, radius = 4
             )

```

***

Enhancements

- Adds in pop-ups to display the body of the tweet

- Removes Stroke and creates 50% opacity

- Changes the radius to 4



### Add in a contour map

```{r}

## MAKE CONTOUR LINES
## Note, bandwidth choice is based on MASS::bandwidth.nrd()
kde <- bkde2D(points[ , c("point_long", "point_lat")],
              bandwidth=c(.0038, .0058), gridsize = c(100,100))
CL <- contourLines(kde$x1 , kde$x2 , kde$fhat)

## EXTRACT CONTOUR LINE LEVELS
LEVS <- as.factor(sapply(CL, `[[`, "level"))
NLEV <- length(levels(LEVS))

## CONVERT CONTOUR LINES TO POLYGONS
pgons <- lapply(1:length(CL), function(i)
    Polygons(list(Polygon(cbind(CL[[i]]$x, CL[[i]]$y))), ID=i))
spgons = SpatialPolygons(pgons)


leaflet(spgons) %>%
  addTiles() %>%
  addPolygons(color = heat.colors(NLEV, NULL)[LEVS]) %>%
  addCircleMarkers(lng=points$point_long, lat=points$point_lat,   popup = points$body, #color = pal(points$generator),
             stroke = FALSE, fillOpacity = 0.5, radius = 4
             )
```

***

This step adds in a contour map (kernel density)

- One problem with the points only are multiple points on a specific location (e.g., Bank of America stadium)

- A contour map helps to understand the point density around a broader area.

- Uptown, South End and NoDa are the most concentrated areas.

- There are smaller areas in South Charlotte, Tega Cay and North Charlotte.



### Add in Marker Clusters

```{r}

leaflet(spgons) %>%
  addTiles() %>%
  addPolygons(color = heat.colors(NLEV, NULL)[LEVS]) %>%
  addCircleMarkers(lng=points$point_long, lat=points$point_lat,   popup = points$body, #color = pal(points$generator),
             stroke = FALSE, fillOpacity = 0.5, radius = 10, clusterOptions = markerClusterOptions()
             )
```

***

Enhancement

- Adds in the markerCluster option

- This allows you to interactively measure the number of tweets in an area.