---
title: "Dashboard DEMO"
date: '`r format(Sys.Date(), format="%B %d, %Y")`'
output:
  flexdashboard::flex_dashboard:
    navbar:
      - { title: "by Andy Grogan-Kaylor", href: "https://agrogan1.github.io/", align: left }
    theme:
      version: 4
      bootswatch: cerulean
    vertical_layout: fill
    orientation: rows
    social: [ "twitter", "facebook", "menu" ]
    source_code: embed
---


```{r}

library(knitr) # knitr

library(flexdashboard) # flexdashboard

library(readr) # to read csv

library(DT) # data table

library(formattable) # formatted data table

library(rpivotTable) # pivot table

library(ggplot2) # nice graphs

library(ggthemes) # themes for ggplot2

library(plotly) # interactive graphs

library(dplyr) # data wrangling

library(leaflet) # interactive maps

```

```{r}

simulated_agency_data <- 
  read_csv("simulated agency data.csv")

simulated_agency_data$gender <- 
  as.factor(simulated_agency_data$gender)

simulated_agency_data$raceEthnicity <- 
  as.factor(simulated_agency_data$raceEthnicity)

simulated_agency_data$program <- 
  as.factor(simulated_agency_data$program)

simulated_agency_data$age <- 
  as.numeric(simulated_agency_data$age)

```

```{r}

mycolors <- c("blue", "#FFC125", "darkgreen", "darkorange")

```

Interactive DataViz
=====================================  

Row
-------------------------------------

### allow agency managers and program managers to get a quick *overview* of client well-being.

```{r}

valueBox(paste("Dashboards"),
         color = "success")

```

### All data are simulated and randomly generated

```{r}

valueBox(paste("Simulation"),
         color = "warning")

```

### Clients

```{r}

valueBox(length(simulated_agency_data$ID), 
         icon = "fa-user",
         color = "success")

```

### **Average Mental Health Score**

```{r}

gauge(round(mean(simulated_agency_data$mentalHealthScore), 
            digits = 2),
      min = 0,
      max= 150,
      gaugeSectors(success = c(100, 150), 
                   warning = c(90, 100), 
                   danger = c(0, 90),
                   colors = c("green", 
                              "yellow", 
                              "red")))

```

### Program A

```{r}

valueBox(sum(simulated_agency_data$program == "Program A"),
         icon = "fa-building-o",
         color = "primary")

```

### Program B

```{r}

valueBox(sum(simulated_agency_data$program == "Program B"),
         icon = "fa-building",
         color = "secondary")

```

### Program C

```{r}

valueBox(sum(simulated_agency_data$program == "Program C"),
         icon = "fa-building-o",
         color = "primary")

```

### Program D

```{r}

valueBox(sum(simulated_agency_data$program == "Program D"),
         icon = "fa-home",
         color = "secondary")

```

Row
-------------------------------------

### Race or Ethnicity

```{r}

p1 <- simulated_agency_data %>%
  group_by(raceEthnicity) %>%
  summarize(count = n()) %>% 
  plot_ly(x = ~raceEthnicity,
              y = ~count, 
              color = I("blue"), 
              type = "bar") %>%
  layout(xaxis = list(title = 'Race or Ethnicity'),
         yaxis = list(title = 'Count'))

p1

```

### Gender

```{r}

p4 <- simulated_agency_data %>%
  group_by(gender) %>%
  summarize(count = n()) %>% 
  plot_ly(labels = ~gender, 
          values = ~count,
          marker = list(colors = mycolors)) %>%
  add_pie(hole = 0.4) %>%
  layout(xaxis = list(zeroline = FALSE, 
                      showline = FALSE, 
                      showticklabels = FALSE, 
                      showgrid = FALSE),
         yaxis = list(zeroline = FALSE, 
                      showline = FALSE, 
                      showticklabels = FALSE, 
                      showgrid = FALSE))

p4

```

Row
-------------------------------------

### Mental Health Scores 

```{r}

p0 <- plot_ly(simulated_agency_data, 
              x = ~rownames(simulated_agency_data),
              y = ~mentalHealthScore,
              text = paste("Client ID:", 
                           simulated_agency_data$ID,
                           "
MH Score:", 
                           simulated_agency_data$mentalHealthScore),
              type = "bar") %>%
  layout(xaxis = list(title = 'Client'),
         yaxis = list(title = 'Mental Health Score'))

p0

```

### Mental Health Score by Age

```{r}

p2 <- plot_ly(simulated_agency_data, x = ~age) %>% 
  add_markers(y = ~mentalHealthScore, 
              text = ~paste("MH Score: ", mentalHealthScore), 
              showlegend = FALSE) %>%
  add_lines(y = ~fitted(loess(mentalHealthScore ~ age)),
            name = "Loess Smoother", 
            color = I("#FFC125"),
            showlegend = TRUE, 
            line = list(width = 5)) %>%
  layout(xaxis = list(title = 'Age'),
         yaxis = list(title = 'Mental Health Score'))

p2

```

### Program

```{r}

p3 <- simulated_agency_data %>%
  group_by(program) %>%
  summarize(count = n()) %>% 
  plot_ly(labels = ~program, 
          values = ~count,
          marker = list(colors = mycolors)) %>% 
  add_pie(hole = 0.4) %>% 
  layout(xaxis = list(zeroline = FALSE, 
                      showline = FALSE, 
                      showticklabels = FALSE, 
                      showgrid = FALSE),
         yaxis = list(zeroline = FALSE, 
                      showline = FALSE, 
                      showticklabels = FALSE, 
                      showgrid = FALSE))

p3

```

Map
=====================================  

###  Client Locations (Simulated and Randomly Generated)

```{r}

client_location_data <- simulated_agency_data

N <- length(client_location_data$ID)

client_location_data$mylatitude <- rnorm(N, 42.25, .1)

client_location_data$mylongitude <- rnorm(N, -83.74, .1)

m <- leaflet(data = client_location_data, 
             height = 500, 
             width = 1000)  %>% 
  setView(lng = mean(client_location_data$mylongitude), 
          lat = mean(client_location_data$mylatitude), 
          zoom = 10) %>% 
  addProviderTiles("Stamen.TonerLite", group = "Toner Lite")

m %>% addCircleMarkers(lng = ~mylongitude, 
                       lat = ~mylatitude,
                       popup = ~paste("client:  ", ID),
                       clusterOptions = markerClusterOptions())

```

Data
=====================================  

```{r}

# define an action rule based upon mental health scores

simulated_agency_data <- 
  simulated_agency_data %>% 
  mutate(action_required = ifelse(mentalHealthScore < 90,
                         TRUE, 
                         FALSE))

```

```{r}

as.datatable(formattable(simulated_agency_data, 
                         list(age = color_bar("orange"),
                              mentalHealthScore = color_bar("cyan"),
                              action_required = color_tile("green",
                                                           "red")),
             caption = "Simulated Data on Clients"))

# datatable(simulated_agency_data,
#           caption = "Simulated Data on Clients",
#           rownames = TRUE, 
#           filter = 'top', 
#           options = list(pageLength = 10))

```

Pivot Table
=====================================  

```{r}

rpivotTable(simulated_agency_data, 
            aggregatorName = "Count",
            rows = "program",
            rendererName = "Bar Chart")

```

Report {data-orientation=columns}
=====================================  

Column {data-width=100}
-------------------------------------

### Clients

```{r}

valueBox(length(simulated_agency_data$ID), 
         icon = "fa-user")

```

### Average Age

```{r}

valueBox(mean(simulated_agency_data$age), 
         icon = "fa-area-chart")

```

### Average Mental Health

```{r}

valueBox(round(mean(simulated_agency_data$mentalHealthScore), 
               digits = 2), 
         icon = "fa-area-chart")

```

Column
-------------------------------------

Auto-Generated Report



* This is a report on `r length(simulated_agency_data$ID)` clients.

* The average age was `r mean(simulated_agency_data$age)`.

* The average mental health score was `r round(mean(simulated_agency_data$mentalHealthScore), digits=2)`.



This report was automatically generated on `r format(Sys.Date(), format="%B %d, %Y")`.



Confidentiality 
=====================================  



Think carefully about the confidentiality of your client data, the degree to which the dashboard might identify specific clients, and the audiences with whom the dashboard might be shared:

1. Generally speaking, charts and graphs preserve client confidentiality.  However, if there is a unique or rare combination of data, such a combination may reveal something about an individual client.

2. Generally speaking, tables of individual data, contain a great potential to identify individual clients, and therefore access to such tables must be carefully thought through.

3. With GIS data, it is particularly important to think critically about data access, as the identifiability issues may not be readily apparent. At a high level overview (i.e. a map that is "zoomed out"), GIS data may not pose a great deal of identifiability risk.  However, GIS data can be **highly identifying**, when zoomed in to the street or block level.