Air-crash Data Visualization
Adwesh Behera
Goal
The article describes the visualization of air-crash data with the help of map. The data is hosted by the National Transportation Safety Board (NTSB) of USA, a specialized federal wing that is assigned the task of investigating accidents in road, air, water and pipeline. Apart from this, it also documents the cases and issues advisories to ensure and enhance transportation safety. The data used here can be downloaded from NTSB official website.
The Data
Below is a summary of the original imported dataset.
## Event.Id Investigation.Type Accident.Number
## 20001212X19172 : 3 Accident :75722 ANC00FA018 : 1
## 20001214X45071 : 3 Incident : 3160 ANC00FA024 : 1
## 20100204X45658 : 3 ANC00FA052 : 1
## 20101022X34140 : 3 ANC00FA056 : 1
## 20001204X00086 : 2 ANC00FA076 : 1
## 20001205X00276 : 2 ANC00FA081 : 1
## (Other) :78866 (Other) :78876
## Event.Date Location Country
## 05/16/1982 : 25 ANCHORAGE. AK : 372 United States :74421
## 06/30/1984 : 25 MIAMI. FL : 185 : 507
## 07/08/2000 : 25 CHICAGO. IL : 169 Canada : 245
## 05/17/1986 : 24 ALBUQUERQUE. NM : 165 Brazil : 213
## 06/05/1983 : 24 HOUSTON. TX : 155 Mexico : 209
## 08/05/1983 : 24 FAIRBANKS. AK : 138 United Kingdom : 208
## (Other) :78735 (Other) :77698 (Other) : 3079
## Latitude Longitude Airport.Code
## Min. :-78.02 Min. :-178.68 :34329
## 1st Qu.: 33.40 1st Qu.:-115.02 NONE : 1462
## Median : 38.17 Median : -94.53 PVT : 354
## Mean : 37.71 Mean : -93.81 ORD : 146
## 3rd Qu.: 42.56 3rd Qu.: -81.75 APA : 141
## Max. : 89.22 Max. : 177.56 MRI : 125
## NA's :53509 NA's :53518 (Other):42325
## Airport.Name Injury.Severity Aircraft.Damage
## :29991 Non-Fatal :59746 : 2395
## N/A : 1669 Fatal(1) : 7770 Destroyed :17242
## PRIVATE : 215 Fatal(2) : 4585 Minor : 2511
## Private : 171 Incident : 3160 Substantial :56734
## NONE : 140 Fatal(3) : 1435
## Private Airstrip : 134 Fatal(4) : 1003
## (Other) :46562 (Other) : 1183
## Aircraft.Category Registration.Number Make
## :56857 : 3003 CESSNA :17004
## Airplane :18875 NONE : 364 PIPER : 9370
## Helicopter : 2317 None : 104 Cessna : 7742
## Glider : 377 USAF : 9 Piper : 4096
## Balloon : 172 N20752 : 8 BEECH : 3113
## Gyrocraft : 101 UNK : 7 Beech : 1748
## (Other) : 183 (Other) :75387 (Other) :35809
## Model Amateur.Built Number.of.Engines
## 152 : 2273 : 574 Min. : 0.000
## 172 : 1241 No :70757 1st Qu.: 1.000
## 172N : 1130 Yes : 7551 Median : 1.000
## PA-28-140 : 900 Mean : 1.148
## 172M : 775 3rd Qu.: 1.000
## 150 : 722 Max. :18.000
## (Other) :71841 NA's :4045
## Engine.Type FAR.Description
## Reciprocating :64202 :56936
## : 3421 Part 91: General Aviation :17632
## Turbo Shaft : 3290 Part 137: Agricultural : 1093
## Turbo Prop : 3010 Non-U.S.. Non-Commercial : 752
## Turbo Fan : 2215 Part 135: Air Taxi & Commuter : 747
## Unknown : 2050 Part 121: Air Carrier : 525
## (Other) : 694 (Other) : 1197
## Schedule Purpose.of.Flight
## :67399 Personal :44337
## NSCH : 3852 Instructional : 9442
## SCHD : 3532 Unknown : 6789
## UNK : 4099 Aerial Application : 4357
## Business : 3860
## : 3803
## (Other) : 6294
## Air.Carrier Total.Fatal.Injuries
## :74980 Min. : 0.000
## UNITED AIRLINES : 49 1st Qu.: 0.000
## AMERICAN AIRLINES : 41 Median : 0.000
## CONTINENTAL AIRLINES : 25 Mean : 0.812
## SOUTHWEST AIRLINES CO : 24 3rd Qu.: 1.000
## USAIR : 24 Max. :349.000
## (Other) : 3739 NA's :23019
## Total.Serious.Injuries Total.Minor.Injuries Total.Uninjured
## Min. : 0.000 Min. : 0.000 Min. : 0.000
## 1st Qu.: 0.000 1st Qu.: 0.000 1st Qu.: 0.000
## Median : 0.000 Median : 0.000 Median : 1.000
## Mean : 0.316 Mean : 0.501 Mean : 5.794
## 3rd Qu.: 0.000 3rd Qu.: 1.000 3rd Qu.: 2.000
## Max. :111.000 Max. :380.000 Max. :699.000
## NA's :25177 NA's :24073 NA's :12126
## Weather.Condition Broad.Phase.of.Flight Report.Status
## : 2127 LANDING :19030 Factual : 325
## IMC : 5647 TAKEOFF :15171 Foreign : 3875
## UNK : 957 CRUISE :10709 Preliminary : 1158
## VMC :70151 MANEUVERING : 9751 Probable Cause :73524
## APPROACH : 7673
## : 6159
## (Other) :10389
## Publication.Date X
## :13389 Mode:logical
## 03/31/1993 : 453 NA's:78882
## 11/25/2003 : 396
## 02/15/2001 : 377
## 09/13/2005 : 325
## 09/14/1993 : 312
## (Other) :63630Filtering Data
The imported data is filtered as per the requirement of visualization. The filter function from dplyr comes in handy here. The data can be filtered with respect to column(s), most of which contain discrete values. This helps make meaningful inferences for various situations. For the purpose of simplicity, all the crash sites with investigation type ‘accident’ and damage ‘destroyed’ are used for visualization. Below is a sample code to filter data.
filter(Accidents, Investigation.Type == " Accident ")
filter(Accidents, Investigation.Type == " Accident ", Aircraft.Damage == " Destroyed ")
filter(Accidents, Investigation.Type == " Incident " | Aircraft.Damage == " Minor ")Plotting Data
The maps are fetched from Google using methods from maps and ggmap package. The map of required location and suitable zoom are imported, onto which the coordinates are plotted. It should be noted that this method eliminates all the non-available data and plots the rest. It should be noted that the imported data belongs to the NTSB database which has a limited jurisdiction outside USA. So the data does not represent all the crash sites, however it represents the crash sites that was investigated by NTSB.
map<- get_map(location = geocode("INDIA"), zoom = 4,maptype = "roadmap", scale = 1)
ggmap(map) +geom_point(data = Accidents, aes(x = Longitude, y = Latitude, fill = "red",alpha = 0.5), size = 4, shape = 21,na.rm = F) +guides(fill=FALSE, alpha=FALSE, size=FALSE)The above is a sample codes for plotting the data. The geocode can be resolved by providing a location string or passing the actual longitude and latitude. Similarly the zoom level can also be set. The maptype should be one of “terrain”, “satellite”, “roadmap”, “hybrid”.
Sample Plots
India
United States of America
Alaska
Anchorage
Africa
Europe
Australia
