Building with Shiny:
Taking R to the Web

 
image/svg+xml
image/svg+xml
image/svg+xml

https://slides.com/alexpawlowski/building-with-shiny/live/

 

29 April 2016

 

Knoxville R Users Group

 

@alexpawlowski

Alex Pawlowski

  • Overview of Shiny
    • Github
  • Background of Clean Power Plan
  • Goal of myCPP
 
  • EIA API
  • myCPP Tool v1 Demo
  • GitBook
  • New!
    • Shiny Modules
  • Shameless Plugs
 

Outline

 

Building with Shiny

taking R to the web

 
  • Framework developed by 
  • take R code and make a web "app" quickly
    • ui.R
      • simple input and output setup
    • server.R
      • reactivity for R 
 

What

 
  • can easily extend/expand

 

 

 

  • multiple tabs
  • control reactivity
  • easy deployment with shinyapps.io, but can run shiny server on web service of choice
1

Capabilities

 
Created by potrace 1.13, written by Peter Selinger 2001-2015 image/svg+xml

Building with Shiny

taking R to the web

 

ui.R

 
Not signed in

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Proin urna odio, aliquam vulputate faucibus id, elementum lobortis felis. Mauris urna dolor, placerat ac sagittis quis.

server.R

 
Not signed in
# ui.R

## Libraries ----
library(shiny)
library(plotly)
library(leaflet)
library(maps)

shinyUI(fluidPage(theme = "bootstrap.css",
                  # Google Analytics
                  tags$head(includeScript("www/google-analytics.js")),
                  # Application title
                  HTML('<img src="logo.svg" style = "max-height:130px" width = "100%"/>'),
                  fluidRow(column(12, align ="center",
                                  h3("A Clean Power Plan Evaluation Tool")
                  )),
# ...

                  fluidRow(
                    selectizeInput("stateInput", #inputID
                                   label = "State", #label
                                   choices = NULL,
                                   selected = "Alabama",
                                   multiple = FALSE,
                                   options = list(placeholder = 'select a state name'))
                  ),
# ...
))
# server.R

## Libraries ----
library(shiny)
library(plotly)
library(leaflet)
library(maps)

## Non-Reactive ---

# ... 
### Plant Location Data
geodata <- read.csv("data/plantgeodata.csv")

# ...

## Reactive ---
shinyServer(function(input, output, session) {
  
  updateSelectizeInput(session,
                       'stateInput',
                       choices = statenames, #must be a character vector!!! http://shiny.rstudio.com/articles/selectize.html
                       selected = "Alabama",
                       server = TRUE)
  
  output$NGCCSlider <- renderUI({
    sliderInput("NGCCCap",
                "NGCC Average Capacity Factor (%)",
                min = 0,
                max = 90,
                value = generationDataCleaned[input$stateInput, "NGCCcapFactor"] * 100,
                step = 1,
                width = '1000px')
    
  })
# ...
  output$ratePlotly <- renderPlotly({
    r <- plot_ly(result(),
                 type = "bar",       # all "bar" attributes: https://plot.ly/r/reference/#bar
                 orientation = "h",
                 x = Rate,               # more about bar's "x": /r/reference/#bar-x
                 y = Name,                # more about bar's "y": /r/reference/#bar-y
                 name = "lbsCO2/MWh",
                 opacity = 0.7,    # more about bar's "name": /r/reference/#bar-name
                 marker = list(         # marker is a named list, valid keys: /r/reference/#bar-marker
                   color=c("517C96","FF8200")     # more about marker's "color" attribute: /r/reference/#bar-marker-color
                 ))
    r <- add_trace(x = c(result()[1,5],result()[1,5]), type = "line",
                   marker = list(
                     color="red",
                     size = 0
                   ),
                   line = list(          # marker is a named list, valid keys: /r/reference/#bar-marker
                     color="red",
                     width = 6
                     
                   ),
                   name = "Goal"
    )
    r <- layout(r,              # all of layout's properties: /r/reference/#layout
                xaxis = list(           # layout's xaxis is a named list. List of valid keys: /r/reference/#layout-xaxis
                  title = ""     # xaxis's title: /r/reference/#layout-xaxis-title
                ),
                yaxis = list(           # layout's yaxis is a named list. List of valid keys: /r/reference/#layout-yaxis
                  title = "lbsCO2/MWh"      # yaxis's title: /r/reference/#layout-yaxis-title
                ),
                margin = list(
                  l = 120
                )
                
    )
    
  })


}) /end reactive

Building with Shiny

taking R to the web

 
  • Leveraging R with Shiny framework, while collaborating with a team
    • opportunity to introduce version control
    • introduce team to open source community
 

Goal

 
  • too many things introduced at a time
    • learning R
    • using RStudio
    • learning Shiny
    • learning git, github
  • "version control" :: emails
    • zip files
    • merge was fun...
1

Reality

 
image/svg+xml

Building with Shiny

taking R to the web

 

Clean Power Plan

 
  • Effort to cut US power plant emissions by 32% from 2005 levels in 2030 with state-specific plans
  • Implementation begins in 2022
  • Provide states flexibility in energy options to meet plan
 

What is it?

 

Timeline

 

Building with Shiny

taking R to the web

 
  • June 2014, proposed
  • Aug  2015, announced
  • Oct  2015, 24 states sue
  • Feb 9, 2016, Supreme Court stayed plan pending review
  • June 2016 - suit to be heard in DC District Court
States that support / deny CPP image/svg+xml States that support / deny CPP Sue EPA Later joined Support EPA Neutral Exempt Hawaii Alaska New Hampshire Vermont Maine Rhode Island New York Pennsylvania Washington, D.C. Florida Michigan California Puerto Rico U.S. Virgin Islands Guam Northern Mariana Islands American Samoa

States' Views on Clean Power Plan

 

Goal of myCPP

 
  • easy to use, intuitive tool for policy stakeholders to understand potential strategies to meet CPP
    • use potential strategy to begin more detailed feasibility
  • help users visualize current energy sources in production and emissions impacts
 

Defined

 
  • Costs of power facilities aren't considered
    • costs can vary per state, incentives as well
    • proprietary concerns
    • wrong numbers can be worse than having none
  • energy from across states not currently considered
1

Limitations

 

Building with Shiny

taking R to the web

 

     Data

 
  • REST API
    • return JSON or XML
    • use plant ID to send request
  • Data of interest
    • location of plant
  • Data not used
    • historical energy production
 

API

 
  • Real data of interest locked away in excel docs (that needed formatting anyhow)
    • Form 860 - Generation
    • Form 923 - Ops
  • Data of interest
    • heat rate of combustion plants
    • energy use / plant / state
    • energy demand
1

Excel forms :(

 
image/svg+xml

Building with Shiny

taking R to the web

 
  • idea to initial setup: couple hours
  • time to v.1
    • a lot of love
    • deadlines
      moved up
      quickly, difficult for best diffusion of work
 

Speed Development

 

Open Source Tool built on Open Source Tools

 
Created by potrace 1.13, written by Peter Selinger 2001-2015 image/svg+xml
image/svg+xml

Benefits

 

Building with Shiny

taking R to the web

 

Building with Shiny

taking R to the web

 
Created by potrace 1.13, written by Peter Selinger 2001-2015
  • great for documentation
  • searchable
  • integrates with github (uses Git)
  • responsive
  • easily hook to custom domain
  • Written in Markdown
  • customize CSS

Gitbook

 

Features

 

Building with Shiny

taking R to the web

 

The Team

 

Alex Pawlowski

 

Justin Knowles

 

Michelle Halsted

 

Jessica Velez

 

Sarah Eichler-Wood

 

Emilio Ramirez

 
image/svg+xml
image/svg+xml

New!

  • Shiny v0.13.0 20 Jan 2016 released a bombshell of updates
    • Gadgets
      • interactive locally-run data analysis

Shiny Modules

  • Aimed to help us reduce complexity within server.R script - "abstract", as well as easily reuse between projects

Building with Shiny

taking R to the web

 
library(shiny)

source("linked_scatter.R")

ui <- fixedPage(
  h2("Module example"),
  linkedScatterUI("scatters"),
  textOutput("summary")
)

server <- function(input, output, session) {
  df <- callModule(linkedScatter, "scatters", reactive(mpg),
    left = reactive(c("cty", "hwy")),
    right = reactive(c("drv", "hwy"))
  )

  output$summary <- renderText({
    sprintf("%d observation(s) selected", nrow(dplyr::filter(df(), selected_)))
  })
}

shinyApp(ui, server)

example from RStudio

app.R

["helper"].R

library(shiny)
library(ggplot2)

linkedScatterUI <- function(id) {
  ns <- NS(id)

  fluidRow(
    column(6, plotOutput(ns("plot1"), brush = ns("brush"))),
    column(6, plotOutput(ns("plot2"), brush = ns("brush")))
  )
}

linkedScatter <- function(input, output, session, data, left, right) {
  # Yields the data frame with an additional column "selected_"
  # that indicates whether that observation is brushed
  dataWithSelection <- reactive({
    brushedPoints(data(), input$brush, allRows = TRUE)
  })

  output$plot1 <- renderPlot({
    scatterPlot(dataWithSelection(), left())
  })

  output$plot2 <- renderPlot({
    scatterPlot(dataWithSelection(), right())
  })

  return(dataWithSelection)
}

scatterPlot <- function(data, cols) {
  ggplot(data, aes_string(x = cols[1], y = cols[2])) +
    geom_point(aes(color = selected_)) +
    scale_color_manual(values = c("black", "#66D65C"), guide = FALSE)
}

Building with Shiny

taking R to the web

 

Building with Shiny

taking R to the web

 

on

Join 330 Developers in Knoxville

28 April, 2nd Quarterly Meetup!

Links

 
Created by potrace 1.13, written by Peter Selinger 2001-2015

mycpp.gitbooks.io/mycpp/content/

 

bccpp.shinyapps.io/mycpp/

 

mycpp

 
image/svg+xml

Copy of Building With Shiny

By Alex Pawlowski

Copy of Building With Shiny

Short talk about lessons learned to lead a team to develop an interactive visualization tool using the Shiny Framework in R.

  • 775