Examining US Cocaine Seizures in 2007

Introduction

• A Shiny app that summarizes and displays data about US cocaine seizures during 2007

• The app presents 3 maps, each showing the following information by US state for a particular month:

  • Average cocaine purity (%)
  • Total seizure weight (grams)
  • Total seizure value (US$)

• All the user has to do is to select a month (in numeric format) using a slider; the maps are automatically updated.

• The app can be found here

• Source code can be found here

What it looks like

• Scroll up and down to see the 3 maps

• The data is included in the ggvis package; it comes from DEA's STRIDE, System to Retrieve Information from Drug Evidence

Dataset Treatment Example 1

• The dataset consists of 3380 observations, each comprising

  • Abbreviation for the state in which the seizure took place
  • Potency of the drug expressed as a percentage
  • Weight of the seizure in grams
  • Month in which the seizure occurred
  • Estimated value of the seizure in USD.

• State abbreviations are converted to to full state names in lowercase

    library(ggvis); data(cocaine)
    convertName <- function(abb) {
       tolower(state.name[grep(abb, state.abb)])
    }     
    cocaine$state <- as.character(sapply(cocaine$state, convertName)) 

Dataset Treatment Example 2

• The data is filtered by month; then an appropriate summary is obtained (mean/total)

      theMonth <- 6  # supplied by the user through the slider!
      prepData1 <- function(theMonth) {
         theDataTable <- data.table(subset(cocaine, month == theMonth))
         avgPurityByState <- theDataTable[,list(meanPurity=mean(potency)),by=state]
      }
      avgPurityByState <- prepData1(theMonth)
      print(head(avgPurityByState))   # let's take a look at the results....

##           state meanPurity
## 1:      indiana   60.26667
## 2:      florida   61.42308
## 3: pennsylvania   45.30000
## 4:   new jersey   44.50000
## 5:     illinois   53.20000
## 6:   california   64.75000