Pre-Load useful R packages.

USA Mining Safety data

This plot draws data from the US government website for the Mine Safety and Health Administrations.

The data for fatalities can be found at: https://arlweb.msha.gov/stats/centurystats/coalstats.asp https://arlweb.msha.gov/stats/centurystats/mnmstats.asp
Legislative events related to miner safety were taken from: https://www.msha.gov/about/history
Data for disasters with more than 5 people killed was taken from: https://wwwn.cdc.gov/NIOSH-Mining/MMWC/MineDisasters/Table

Fatality Rates

Read in the data for Miner fatalities from 1900 to present for both Coal mines and Metal/NonMetal Mines. Calculate the fatality rate per 1000 miners employed that year to account for decrease in the number of workers in this industry.

mines <- read.csv("miners.csv")
mines %<>% mutate(MiningSector=as.factor(MiningSector))
mines <- mines %>% mutate(Miners=str_replace(Miners,",",""),Fatalities=str_replace(Fatalities,",",""),Miners=str_replace(Miners,",",""),Fatalities=str_replace(Fatalities,",","")) %>% mutate(Miners=as.integer(Miners),Fatalities=as.integer(Fatalities))

mines %<>% mutate(perworkerfatality=1000.0*Fatalities/Miners)
#missing data for Metal and Mon Metal mines
# mines %>% filter(is.na(Miners))

Read in a list of major actions by the federal government related to mine safety.

minehistory <- read.csv("MiningSafetyHistory.csv")

Create a plot that shows the rate of fatalities for coal mines versus non-coal mines (ie. metal, stone) and labels important historic points. The 1969 legislation is highlighted because it is considered an important step because it created criminal and monitary punishments for safety violators.

p_fatalities <- ggplot()+
  geom_line(data=mines,mapping=aes(x=Year,y=perworkerfatality,
                                   color=MiningSector),size=1.25)+
  theme_classic()+
  geom_textvline(data=minehistory %>% filter(Year!=1969),
                 mapping=aes(xintercept=Year,label=Event),
                 color="darkgrey",size=2.62,linetype="dotted",hjust=1)+
  geom_textvline(data=minehistory %>% filter(Year==1969),
                 mapping=aes(xintercept=Year,label=Event),
                 color="red",size=2.62,linetype="dotted",hjust=1)

eventyear <- c(minehistory$Year,1900,2021)
p_fatalies_final <- p_fatalities+
  theme(legend.position=c(0.2,.08), 
        legend.title = element_blank(),
        legend.background = element_blank(),
        axis.text.x = element_text(angle = 90,vjust=0.5))+
  ylab("Fatalities per 1000 Workers")+
  scale_x_continuous(breaks=eventyear)+
  coord_cartesian(ylim=c(0.15,5.5),xlim=c(1905,2018))+
  scale_color_discrete(labels=c("Coal Mines","Metallic/NonMetallic Mines"))+
  ggtitle("A History of Mine Fatality Rates in the USA since 1900")
p_fatalies_final

## Warning: Removed 31 row(s) containing missing values (geom_path).

ggsave("mine_fatalities.png",plot=p_fatalies_final)

## Saving 7 x 5 in image

## Warning: Removed 31 row(s) containing missing values (geom_path).

Types of Mining Disasters

Let’s also look at the types of mining accidents and whether their distribution among categories has changed since the sixties. The data is filtered from 1900 to present, giving about 70 years of data before and after the legislation.

disasters <- read.csv("MiningDisasters.csv")
disasters %<>% mutate(Mining.Sector=str_to_lower(Mining.Sector))
disasters %<>% mutate(Date=parse_date_time(Date,"mdy")) %>% mutate(Year=year(Date))
disasters %<>% mutate(across(c(Mine.Type,State,City,Mining.Sector,Mine.Type),as.factor))
disasters %>% head()

##         Date              Mine.Name         City State Killed Product
## 1 1839-03-17            Black Heath     Richmond    VA     53    Coal
## 2 1844-06-14            Black Heath     Richmond    VA     11    Coal
## 3 1846-01-11                  No. 1   Carbondale    PA     14    Coal
## 4 1847-02-18                Spencer   Pottsville    PA      7    Coal
## 5 1850-11-29 Cox's Pit, Clover Hill   Winterpock    VA      7    Coal
## 6 1854-05-14           Chesterfield New Richmond    VA     20    Coal
##   Accident.Type Mine.Type Mining.Sector Year
## 1     Explosion        UG          coal 1839
## 2     Explosion        UG          coal 1844
## 3       Cave-in        UG          coal 1846
## 4     Explosion        UG          coal 1847
## 5     Explosion        UG          coal 1850
## 6     Explosion        UG          coal 1854

Using String Searches, let’s try to lump Accident Types into general categories based on those named in https://wwwn.cdc.gov/NIOSH-Mining/MMWC/MineDisasters/AccidentType#

disasters %<>% mutate(AccidentCategory=NA)
disasters %<>% mutate(Accident.Type=str_to_lower(Accident.Type)) 

# classify with categories like https://wwwn.cdc.gov/NIOSH-Mining/MMWC/MineDisasters/AccidentType#
#can get a list of all accident types
# x <- unique(disasters$Accident.Type)
#can test searching for different string patterns in this list
# x[grepl("roof",x)]


disasters %<>% mutate(AccidentCategory=ifelse(grepl('explosion|detonat|blast|explosives|exposive',disasters$Accident.Type),"Explosion",AccidentCategory)) 
disasters %<>% mutate(AccidentCategory=ifelse(grepl('fire',disasters$Accident.Type),"Fires",AccidentCategory))
disasters %<>% mutate(AccidentCategory=ifelse(grepl('flood|water|inund|dam fail|snow',disasters$Accident.Type),"Water",AccidentCategory))
 disasters %<>% mutate(AccidentCategory=ifelse(grepl('suff|smoke|fumes|sulfide gas|inrush hot gases',disasters$Accident.Type),"Suffocation",AccidentCategory))
disasters %<>% mutate(AccidentCategory=ifelse(grepl('haul|hoist|drawbar|cage|man-car|crosshead',disasters$Accident.Type),"Machinery/Equipment fall",AccidentCategory))
disasters %<>% mutate(AccidentCategory=ifelse(grepl('cave-in|roof|fall of|fall and slide|slide of bank|snow|collapse load',disasters$Accident.Type),"Ground Fall",AccidentCategory))
# disasters %<>% mutate(AccidentCategory=ifelse(grepl('bump',disasters$Accident.Type),"Bump/Collapse with Bump",AccidentCategory))
disasters %<>% mutate(AccidentCategory=as.factor(AccidentCategory))
disasters %>% select(Accident.Type,AccidentCategory) %>% head()

##   Accident.Type AccidentCategory
## 1     explosion        Explosion
## 2     explosion        Explosion
## 3       cave-in      Ground Fall
## 4     explosion        Explosion
## 5     explosion        Explosion
## 6     explosion        Explosion

Keep the type of mine consistent with our earlier plot.

disasters %<>% mutate(CoalNonCoal=ifelse(disasters$Mining.Sector=='coal',"Coal Mines","Other Mines"))

Divide the data to times before and after the major safety legistation of the 60s.

disasters %<>% mutate(beforeCoalAct=cut(year(Date),breaks=c(1800.,1969.,2022.)))
disasters %<>% mutate(beforeMetalAct=cut(year(Date),breaks=c(1800.,1966.,2022.)))

Plot the distributions of accident types

disasters %>% head()

##         Date              Mine.Name         City State Killed Product
## 1 1839-03-17            Black Heath     Richmond    VA     53    Coal
## 2 1844-06-14            Black Heath     Richmond    VA     11    Coal
## 3 1846-01-11                  No. 1   Carbondale    PA     14    Coal
## 4 1847-02-18                Spencer   Pottsville    PA      7    Coal
## 5 1850-11-29 Cox's Pit, Clover Hill   Winterpock    VA      7    Coal
## 6 1854-05-14           Chesterfield New Richmond    VA     20    Coal
##   Accident.Type Mine.Type Mining.Sector Year AccidentCategory CoalNonCoal
## 1     explosion        UG          coal 1839        Explosion  Coal Mines
## 2     explosion        UG          coal 1844        Explosion  Coal Mines
## 3       cave-in        UG          coal 1846      Ground Fall  Coal Mines
## 4     explosion        UG          coal 1847        Explosion  Coal Mines
## 5     explosion        UG          coal 1850        Explosion  Coal Mines
## 6     explosion        UG          coal 1854        Explosion  Coal Mines
##        beforeCoalAct     beforeMetalAct
## 1 (1.8e+03,1.97e+03] (1.8e+03,1.97e+03]
## 2 (1.8e+03,1.97e+03] (1.8e+03,1.97e+03]
## 3 (1.8e+03,1.97e+03] (1.8e+03,1.97e+03]
## 4 (1.8e+03,1.97e+03] (1.8e+03,1.97e+03]
## 5 (1.8e+03,1.97e+03] (1.8e+03,1.97e+03]
## 6 (1.8e+03,1.97e+03] (1.8e+03,1.97e+03]

disasters %>% filter(year(Date)>=1900) %>% group_by(CoalNonCoal) %>% summarise(z=sum(Killed))

## # A tibble: 2 × 2
##   CoalNonCoal     z
##   <chr>       <int>
## 1 Coal Mines  11703
## 2 Other Mines  1149

noncoaldisaster <- disasters %>% filter(year(Date)>=1900,CoalNonCoal=="Other Mines") 
noncoaldisaster<- noncoaldisaster%>% group_by(beforeMetalAct,AccidentCategory) %>% summarise(Deaths=sum(Killed)) %>% ungroup() %>% complete(beforeMetalAct,AccidentCategory,fill=list(N=0,Deaths=0))

## `summarise()` has grouped output by 'beforeMetalAct'. You can override using
## the `.groups` argument.

coaldisaster <- disasters %>% filter(year(Date)>=1900,CoalNonCoal=="Coal Mines")
coaldisaster<- coaldisaster %>% group_by(beforeCoalAct,AccidentCategory) %>% summarise(Deaths=sum(Killed)) %>% ungroup() %>% complete(beforeCoalAct,AccidentCategory,fill=list(N=0,Deaths=0))

## `summarise()` has grouped output by 'beforeCoalAct'. You can override using the
## `.groups` argument.

p_coal_accidents <- ggplot(data=coaldisaster %>% ungroup() %>% complete()
       )+geom_col(mapping=aes(x=AccidentCategory,y=Deaths,fill=beforeCoalAct),position="dodge")+
  ggtitle("Coal Mine Disasters from 1900 to 2021")+
  theme_classic()+
  theme(legend.position=c(0.65,.90), 
        legend.title = element_blank(),
        legend.background = element_blank(),
        axis.text.x = element_text(angle = -10,hjust=0.5,vjust=0.0))+
  ylab("Total Fatalities")+xlab("Accident Type")+
  scale_fill_discrete(labels=c("Before 1969 Coal Mine Health and Safety Act","After Act"))

p_metal_accidents <- ggplot(data=noncoaldisaster %>% ungroup() %>% complete()
       )+geom_col(mapping=aes(x=AccidentCategory,y=Deaths,fill=beforeMetalAct),position="dodge")+
    ggtitle("Metallic/NonMetallic Mine Disasters from 1900 to 2021")+
  theme_classic()+
  theme(legend.position=c(0.65,.90), 
        legend.title = element_blank(),
        legend.background = element_blank(),
        axis.text.x = element_text(angle = -10,hjust=0.5,vjust=0.0))+
  ylab("Total Fatalities")+xlab("Accident Type")+
  scale_fill_discrete(labels=c("Before 1966 Metal and Nonmetallic Mine Safety Act","After Act"))

p_metal_accidents

p_coal_accidents

ggsave("coal_mine_disasters.png",plot=p_coal_accidents)

## Saving 7 x 5 in image

ggsave("non_coal_mine_disasters.png",plot=p_metal_accidents)

## Saving 7 x 5 in image

blankPlot <- ggplot()+geom_blank(aes(1,1))+
  theme(
    plot.background = element_blank(), 
   panel.grid.major = element_blank(),
   panel.grid.minor = element_blank(), 
   panel.border = element_blank(),
   panel.background = element_blank(),
   axis.title.x = element_blank(),
   axis.title.y = element_blank(),
   axis.text.x = element_blank(), 
   axis.text.y = element_blank(),
   axis.ticks = element_blank(),
   axis.line = element_blank()
     )

library(ggpubr)
combinedplot <- grid.arrange(p_fatalies_final,p_coal_accidents+rremove("xlab"),p_metal_accidents+rremove("xlab"),nrow=3,heights=c(2,1,1))

## Warning: Removed 31 row(s) containing missing values (geom_path).

ggsave("Homework1Figure.png",combinedplot,width=6,height=10)

Final Figure

finalimage

566Homework1

Anne Goj

2022-09-20

USA Mining Safety data

Fatality Rates

Types of Mining Disasters

Final Figure