The City University of New York School of Professional Studies Knowledge and Visual Analytics (DATA 608)
Data Insights to Save the life of U.S. Military Personnel
Alexis Mekueko
12/15/2021
Benefits
The interest of exploring U.S. Military death data is visualize these death to the public so that something can be done to reduce this death. We all know that U.S. Military involves politicians, technology, industry, healthcare and government. Thus, by displaying this data to the public, all these entities can contribute each at their power level to take major decisions that could end up saving more lives in military. These decisions can be to improving military mechanics, to helping politicians to make better policy, to adjusting military strategy, to doctors and paramedical to rethink and find appropriate health-plan for military personnel. I plan to become a consultant using my skills as data scientist in various domain of the society to present meaningful report to government entities, companies, and organizations to help them in decision making. So, this project will contribute to building skills necessary for one to be successful in data science.Research questions
What is the death rate of military personnel over the course of 20 years?
What is the death rate of military personnel in active duty of the course of 20 years?
What is the death ration of military personnel by accident and illness?
Do military personnel dies more by homicide than combat?
Do military personnel die more by illness than accident?Data source
We were looking at open-source data like kaggle.com and found some interesting dataset about military that no one has not made a any contribution on it. The original source of the dataset ('ActiveDutyDeathNo') is from: Defense Casualty Analysis System (DCAS) , https://dcas.dmdc.osd.mil/dcas/pages/report_by_year_manner.xhtml. Data is completely free and represents 20 years (1980-2010) of data collected on U.S. Active Duty Military Deaths. The details of the dataset can be seen below:Loading Data
-The dataset is pull out from Github in a csv file into Rstudio. We will use R programming language to manipulate and visualize the dataset.
-In addition, will explore the possibility to use python programming language to build a shiny app.Scrubing Data
Tidying up data
## 'data.frame': 31 obs. of 14 variables:
## $ year : int 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 ...
## $ active.duty : int 2050758 2093032 2112609 2123909 2138339 2150379 2177845 2166611 2121659 2112128 ...
## $ full.time..est..guard.reserve: int 22000 22000 41000 49000 55000 64000 69000 71000 72000 74200 ...
## $ selected.reserve.fte : int 86872 91719 97458 100455 104583 108806 113010 115086 115836 117056 ...
## $ total.military.fte : int 2159630 2206751 2251067 2273364 2297922 2323185 2359855 2352697 2309495 2303384 ...
## $ total.deaths : int 2392 2380 2319 2465 1999 2252 1984 1983 1819 1636 ...
## $ accident : num 1556 1524 1493 1413 1293 ...
## $ hostile.action : int 0 0 0 18 1 0 2 37 0 23 ...
## $ homicide : int 174 145 108 115 84 111 103 104 90 58 ...
## $ illness : int 419 457 446 419 374 363 384 383 321 294 ...
## $ pending : int 0 0 0 0 0 0 0 0 0 0 ...
## $ self.inflicted : int 231 241 254 218 225 275 269 260 285 224 ...
## $ terrorist.attack : int 1 0 2 263 6 5 0 2 17 0 ...
## $ undetermined : int 11 13 16 19 16 22 27 25 26 37 ...##Organizing data
-Checking for missing values
-Checking for empty values##
## The dataset contains missing values for a total record of : 0##
## The dataset contains empty values for a total record of : FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSE FALSEVisualizing Empty and Missing Values
Exploring Data
-Inspect data and understand the characteristic of the data Looking for relationship, patterns and values, ##
## Attaching package: 'gridExtra'## The following object is masked from 'package:dplyr':
##
## combine##
## Attaching package: 'kableExtra'## The following object is masked from 'package:dplyr':
##
## group_rows| year | active.duty | full.time..est..guard.reserve | selected.reserve.fte | total.military.fte | total.deaths | accident | hostile.action | homicide | illness | pending | self.inflicted | terrorist.attack | undetermined |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 1980 | 2050758 | 22000 | 86872 | 2159630 | 2392 | 1556 | 0 | 174 | 419 | 0 | 231 | 1 | 11 |
| 1981 | 2093032 | 22000 | 91719 | 2206751 | 2380 | 1524 | 0 | 145 | 457 | 0 | 241 | 0 | 13 |
| 1982 | 2112609 | 41000 | 97458 | 2251067 | 2319 | 1493 | 0 | 108 | 446 | 0 | 254 | 2 | 16 |
| 1983 | 2123909 | 49000 | 100455 | 2273364 | 2465 | 1413 | 18 | 115 | 419 | 0 | 218 | 263 | 19 |
| 1984 | 2138339 | 55000 | 104583 | 2297922 | 1999 | 1293 | 1 | 84 | 374 | 0 | 225 | 6 | 16 |
| 1985 | 2150379 | 64000 | 108806 | 2323185 | 2252 | 1476 | 0 | 111 | 363 | 0 | 275 | 5 | 22 |
| 1986 | 2177845 | 69000 | 113010 | 2359855 | 1984 | 1199 | 2 | 103 | 384 | 0 | 269 | 0 | 27 |
| 1987 | 2166611 | 71000 | 115086 | 2352697 | 1983 | 1172 | 37 | 104 | 383 | 0 | 260 | 2 | 25 |
| 1988 | 2121659 | 72000 | 115836 | 2309495 | 1819 | 1080 | 0 | 90 | 321 | 0 | 285 | 17 | 26 |
| 1989 | 2112128 | 74200 | 117056 | 2303384 | 1636 | 1000 | 23 | 58 | 294 | 0 | 224 | 0 | 37 |
| 1990 | 2046806 | 74250 | 137268 | 2258324 | 1507 | 880 | 0 | 74 | 277 | 0 | 232 | 1 | 43 |
| 1991 | 1943937 | 70250 | 184002 | 2198189 | 1787 | 931 | 147 | 112 | 308 | 0 | 256 | 0 | 33 |
| 1992 | 1773996 | 67850 | 111491 | 1953337 | 1293 | 676 | 0 | 109 | 252 | 0 | 238 | 1 | 17 |
| 1993 | 1675269 | 68500 | 105768 | 1849537 | 1213 | 632 | 0 | 86 | 221 | 0 | 236 | 29 | 9 |
| 1994 | 1581649 | 65000 | 99833 | 1746482 | 1075 | 544 | 0 | 83 | 206 | 0 | 232 | 0 | 10 |
| 1995 | 1502343 | 65000 | 94585 | 1661928 | 1040 | 538 | 0 | 67 | 174 | 0 | 250 | 7 | 4 |
| 1996 | 1456266 | 65000 | 92409 | 1613675 | 974 | 527 | 1 | 52 | 173 | 0 | 188 | 19 | 14 |
| 1997 | 1418773 | 65000 | 94609 | 1578382 | 817 | 433 | 0 | 42 | 170 | 0 | 159 | 0 | 13 |
| 1998 | 1381034 | 65000 | 92536 | 1538570 | 827 | 445 | 0 | 26 | 174 | 0 | 165 | 3 | 14 |
| 1999 | 1367838 | 65000 | 93104 | 1525942 | 796 | 439 | 0 | 38 | 154 | 0 | 150 | 0 | 15 |
| 2000 | 1372352 | 65000 | 93078 | 1530430 | 832 | 429 | 0 | 37 | 180 | 0 | 153 | 17 | 16 |
| 2001 | 1384812 | 65000 | 102284 | 1552096 | 943 | 461 | 12 | 49 | 197 | 0 | 153 | 46 | 25 |
| 2002 | 1411200 | 66000 | 149942 | 1627142 | 1051 | 565 | 17 | 54 | 213 | 0 | 174 | 0 | 28 |
| 2003 | 1423348 | 66000 | 243284 | 1732632 | 1399 | 597 | 312 | 46 | 231 | 1 | 190 | 0 | 22 |
| 2004 | 1411287 | 66000 | 234629 | 1711916 | 1847 | 605 | 735 | 46 | 256 | 0 | 197 | 0 | 8 |
| 2005 | 1378014 | 66000 | 220000 | 1664014 | 1929 | 646 | 739 | 54 | 280 | 1 | 182 | 0 | 27 |
| 2006 | 1371533 | 72000 | 168000 | 1611533 | 1882 | 561 | 769 | 47 | 257 | 8 | 213 | 0 | 27 |
| 2007 | 1368226 | 72000 | 168000 | 1608226 | 1953 | 561 | 847 | 52 | 237 | 22 | 211 | 0 | 23 |
| 2008 | 1402227 | 73000 | 207917 | 1683144 | 1440 | 506 | 352 | 47 | 244 | 6 | 259 | 1 | 25 |
| 2009 | 1421668 | 75000 | 144083 | 1640751 | 1515 | 467 | 346 | 77 | 277 | 19 | 302 | 0 | 27 |
| 2010 | 1430985 | 76000 | 178193 | 1685178 | 1485 | 424 | 456 | 39 | 238 | 22 | 289 | 0 | 17 |
##
##
## | |usM (N = 31) |
## |:---------------------------------|:--------------------------------------|
## |**year** | |
## | minimum |1,980 |
## | median (IQR) |1,995 (1,987.50, 2,002.50) |
## | mean (sd) |1,995.00 ± 9.09 |
## | maximum |2,010 |
## |**active.duty** | |
## | minimum |1,367,838 |
## | median (IQR) |1,502,343 (1,406,713.50, 2,102,580.00) |
## | mean (sd) |1,702,284.90 ± 337,422.87 |
## | maximum |2,177,845 |
## |**full.time..est..guard.reserve** | |
## | minimum |22,000 |
## | median (IQR) |66,000 (65,000.00, 71,500.00) |
## | mean (sd) |63,614.52 ± 13,263.52 |
## | maximum |76,000 |
## |**selected.reserve.fte** | |
## | minimum |86,872 |
## | median (IQR) |111,491 (96,033.50, 158,971.00) |
## | mean (sd) |131,157.94 ± 46,394.45 |
## | maximum |243,284 |
## |**total.military.fte** | |
## | minimum |1,525,942 |
## | median (IQR) |1,732,632 (1,620,408.50, 2,254,695.50) |
## | mean (sd) |1,897,057.35 ± 318,690.92 |
## | maximum |2,359,855 |
## |**total.deaths** | |
## | minimum |796 |
## | median (IQR) |1,515 (1,063.00, 1,968.00) |
## | mean (sd) |1,575.29 ± 526.56 |
## | maximum |2,465 |
## |**accident** | |
## | minimum |424.00 |
## | median (IQR) |605.00 (516.50, 1,126.00) |
## | mean (sd) |808.81 ± 391.40 |
## | maximum |1,556.00 |
## |**hostile.action** | |
## | minimum |0 |
## | median (IQR) |1 (0.00, 229.50) |
## | mean (sd) |155.29 ± 272.07 |
## | maximum |847 |
## |**homicide** | |
## | minimum |26 |
## | median (IQR) |67 (47.00, 103.50) |
## | mean (sd) |75.13 ± 35.19 |
## | maximum |174 |
## |**illness** | |
## | minimum |154 |
## | median (IQR) |256 (209.50, 342.00) |
## | mean (sd) |276.74 ± 89.15 |
## | maximum |457 |
## |**pending** | |
## | minimum |0 |
## | median (IQR) |0 (0.00, 0.00) |
## | mean (sd) |2.55 ± 6.40 |
## | maximum |22 |
## |**self.inflicted** | |
## | minimum |150 |
## | median (IQR) |231 (189.00, 255.00) |
## | mean (sd) |222.94 ± 43.04 |
## | maximum |302 |
## |**terrorist.attack** | |
## | minimum |0 |
## | median (IQR) |1 (0.00, 5.50) |
## | mean (sd) |13.55 ± 47.44 |
## | maximum |263 |
## |**undetermined** | |
## | minimum |4 |
## | median (IQR) |19 (14.00, 26.50) |
## | mean (sd) |20.29 ± 8.82 |
## | maximum |43 |
Looking at the dataset, we can say the dataset is set in 2 parts:
-Military Personnel = Active.Duty + Full-Time (est.)Guard-Reserve + Selected.Reserve FTEa + Total.Military.FTE.
-Casualty or Type of Death = Total.Deaths + Accident + Hostile.Action + Homicide + Illness + Pending + Self-Inflicted + Terrorist.Attack + Undetermined
-Definition of death rate: the ratio between deaths and individuals in a specified population during a particular time period :
-The incidence of deaths in a given population during a defined time period (such as one year) that is typically expressed per 1000 or 100,000 individuals.
-Total.Death = Total.Deaths + Accident + Hostile.Action + Homicide + Illness + Pending + Self-Inflicted + Terrorist.Attack + Undetermined
-Total.Military.FTE = Active.Duty + Full-Time (est.)Guard-Reserve + Selected.Reserve FTEa
-Death rate per year = (Total.Death/Total.Military.FTE)*100000
-Death rate% per year = (Total.Death/Total.Military.FTE)*100
-Growth rate in total personnel% per year = (Total.Military.FTE(next_year)-Total.Military.FTE(current_year))*100
-Growth rate in total death % per year = (Total.Death(next_year)-Total.Death(current_year))*100Modelind Data
-Exploring other data visualization charts
-Explore building apps to display plots: Shiny, Dash
##
## Attaching package: 'reshape2'## The following object is masked from 'package:tidyr':
##
## smiths
Another way to plot the military personnel against time is to group time block (1980 to 1990 = decade1, 1990 to 2000= decade2, 2000 to 2010 = decade3), while summing the values of other variables within decade. Then, use barplot() or bubble plot. We notice that there is discrepancy active.duty, total.military.fte and full.time..est..guard.reserve,selected.reserve.fte . We can fix this by plotting the 02 variables seperately.
## NOTE: Either Arial Narrow or Roboto Condensed fonts are required to use these themes.## Please use hrbrthemes::import_roboto_condensed() to install Roboto Condensed and## if Arial Narrow is not on your system, please see https://bit.ly/arialnarrow## Loading required package: viridisLite
Let’s see the military death over the 20 years(1980-2010)
##
## Attaching package: 'corrgram'## The following object is masked from 'package:plyr':
##
## baseball## The following object is masked from 'package:lattice':
##
## panel.fill
##
## Attaching package: 'plotly'## The following objects are masked from 'package:plyr':
##
## arrange, mutate, rename, summarise## The following object is masked from 'package:lessR':
##
## style## The following object is masked from 'package:ggplot2':
##
## last_plot## The following object is masked from 'package:stats':
##
## filter## The following object is masked from 'package:graphics':
##
## layoutTotal.Death = Total.Deaths + Accident + Hostile.Action + Homicide + Illness + Pending + Self-Inflicted + Terrorist.Attack + Undetermined Total.Military.FTE = Active.Duty + Full-Time (est.)Guard-Reserve + Selected.Reserve FTEa
Let’s see the total death by year over the course of the 20 years 
Some Reference in U.S. Military Battlefield for the past decades:
1983 - U.S. Military in Grenada
1989 - U.S. Military in Panama
1990 - U.S. Military in Gulf War
1993 - U.S. Military in Somalia War
2001- U.S in Afghanistan (2001-2021)
2003 - U.S. Military in U.S. Iraq(2003-2011)
Let’s Visualize the differente rates among military personnel.
## -- Attaching packages --------------------------------------- tidyverse 1.3.1 --## v tibble 3.0.6 v stringr 1.4.0
## v readr 1.4.0 v forcats 0.5.1
## v purrr 0.3.4## -- Conflicts ------------------------------------------ tidyverse_conflicts() --
## x plotly::arrange() masks plyr::arrange(), dplyr::arrange()
## x gridExtra::combine() masks dplyr::combine()
## x purrr::compact() masks plyr::compact()
## x plyr::count() masks dplyr::count()
## x Matrix::expand() masks tidyr::expand()
## x plyr::failwith() masks dplyr::failwith()
## x plotly::filter() masks dplyr::filter(), stats::filter()
## x kableExtra::group_rows() masks dplyr::group_rows()
## x plyr::id() masks dplyr::id()
## x dplyr::lag() masks stats::lag()
## x purrr::lift() masks caret::lift()
## x plotly::mutate() masks plyr::mutate(), dplyr::mutate()
## x Matrix::pack() masks tidyr::pack()
## x arules::recode() masks lessR::recode(), dplyr::recode()
## x plotly::rename() masks plyr::rename(), dplyr::rename()
## x plotly::summarise() masks plyr::summarise(), dplyr::summarise()
## x plyr::summarize() masks dplyr::summarize()
## x Matrix::unpack() masks tidyr::unpack()
## >>> Suggestions
## PieChart(death.cause, hole=0) # traditional pie chart
## PieChart(death.cause, values="%") # display %'s on the chart
## BarChart(death.cause) # bar chart
## Plot(death.cause) # bubble plot
## Plot(death.cause, values="count") # lollipop plot
##
## --- death.rate ---
##
## n miss mean sd min mdn max
## 8 0 12.500 22.152 0.000 3.865 65.050Interpret Results
- There are 03 set of high correlation among the cause of death in military personnel from 1980 to 2010.
- Illness and Homicide
- Illness and Accident
- Homicide and Accident
- These correlations show that military personnel death by accident and homicide increase with those dying based on illness. In order, the more military personnel are sick, the likely-hood of more death occurring by accident and homicide.
- These correlations shows also that there is likely more death to occur by homicide when more military personnel die by accident.
- U.S. Military tends to have more casualty when engaging in war.
- Over the course of 20 years(1980-2010) of active duty, U.S. Military has significantly dropped. This might have some explanation with United Nations policy on regulating the size of military of countries around the world.Challenges
-There are few challenges in this project to be overcome:
-Due to the sensibility of the dataset, it can be though to be neutral.
-Rendering the data to a suitable chart was not easy.
-There were a confusion perhaps not in the sense of grammar but more of statistical appreciation of what growth rate is. We thought some information could be reveal while exploring the growth rate among total death in U.S. Military personnel. We wanted to see if this rate was increasing or decreasing from year to year. In addition, let’s say a virus is spraying in a population, the rate at which the population is getting contaminated starts at 0 (meaning there were no precedent of such a virus). Therefore, it makes sense to have this rate to eventually settle around zero when the virus is under control and the population is immunized. If we consider this assumption, therefore, it makes no sense to see throughout the timeline of U.S. battlefield the growth rate in total death of military personnel to goes negative. We suspect the formula may need a closer look.
References
https://www.codegrepper.com/code-examples/whatever/insert+image+in+r+markdown
https://sgp.fas.org/crs/natsec/RL32492.pdf
https://cran.r-project.org/web/packages/kableExtra/vignettes/awesome_table_in_html.html
https://www.r-graph-gallery.com/web-line-chart-with-labels-at-end-of-line.html
https://www.r-graph-gallery.com/37-barplot-with-number-of-observation.html
https://sgp.fas.org/crs/natsec/RL32492.pdf
https://bookdown.org/chua/ber642_advanced_regression/r-basics.html