# Step 1: Load R Packages
### options(repos='http://cran.rstudio.com/')
#install.packages("astsa")
#install.packages('ggplot2')
#install.packages('forecast')
#install.packages('tseries')
#install.packages("data.table")
library(astsa)
library(forecast)
library(tseries)
library(zoo)
library(tseries)
library(data.table)
library(dplyr)
library(tidyr)
library(naniar)
library(stringr)
library(ggplot2)
library(DT)
library(lubridate)
library(ggpubr)
setwd("/scratch/user/cma16/Task4_Deliverable2/Process4/AllCrash/FacilityBased/")
load("./multi-lane_divided_WA_reduce_withCrash.rData")
mytype = 'RMD'
setwd(paste0("/scratch/user/cma16/Task4_Deliverable2/Process4/AllCrash/FacilityBased/",mytype))
df_RMD <- W_mun_med
dim(df_RMD)
## [1] 3784320 64
df_RMD$spd_av = 3600*df_RMD$DISTANCE/df_RMD$Travel_TIME_ALL_VEHICLES
df_RMD$spd_pv = 3600*df_RMD$DISTANCE/df_RMD$Travel_TIME_PASSENGER_VEHICLES
df_RMD$spd_ft = 3600*df_RMD$DISTANCE/df_RMD$Travel_TIME_FREIGHT_TRUCKS
### Month, Day
df_RMD$date <- as.character(df_RMD$DATE)
df_RMD$date <- str_pad(df_RMD$DATE, 8, pad = "0")
df_RMD$Month <- substr(df_RMD$date, start = 1, stop = 2)
df_RMD$Day <- substr(df_RMD$date, start = 3, stop = 4)
df_RMD$Year <- substr(df_RMD$date, start = 5, stop = 8)
df_RMD$MonthDay <- paste0(df_RMD$Month,"_", df_RMD$Day)
head(df_RMD)
## TimeStamp TMC V1 DATE EPOCH15
## 1: 114N05428_0101_0 114N05428 425569 1012015 0
## 2: 114N05428_0101_1 114N05428 425570 1012015 1
## 3: 114N05428_0101_10 114N05428 425579 1012015 10
## 4: 114N05428_0101_11 114N05428 425580 1012015 11
## 5: 114N05428_0101_12 114N05428 425581 1012015 12
## 6: 114N05428_0101_13 114N05428 425582 1012015 13
## Travel_TIME_ALL_VEHICLES Travel_TIME_PASSENGER_VEHICLES
## 1: NA NA
## 2: NA NA
## 3: NA NA
## 4: NA NA
## 5: NA NA
## 6: NA NA
## Travel_TIME_FREIGHT_TRUCKS NP ADMIN_LEVE ADMIN_LE_1 ADMIN_LE_2 DISTANCE
## 1: NA N USA Washington King 8.3187
## 2: NA N USA Washington King 8.3187
## 3: NA N USA Washington King 8.3187
## 4: NA N USA Washington King 8.3187
## 5: NA N USA Washington King 8.3187
## 6: NA N USA Washington King 8.3187
## ROAD_NUMBE ROAD_NAME LATITUDE LONGITUDE ROAD_DIREC ORN_FID FID_1
## 1: US-2 47.71281 -121.2902 Westbound 2638.708 1096.573
## 2: US-2 47.71281 -121.2902 Westbound 2638.708 1096.573
## 3: US-2 47.71281 -121.2902 Westbound 2638.708 1096.573
## 4: US-2 47.71281 -121.2902 Westbound 2638.708 1096.573
## 5: US-2 47.71281 -121.2902 Westbound 2638.708 1096.573
## 6: US-2 47.71281 -121.2902 Westbound 2638.708 1096.573
## ACCESS LSHL_TY2 LSHL_TYP MED_TYPE NHS_IND PRK_ZNE RSHL_TY2 RSHL_TYP
## 1: P A Y A
## 2: P A Y A
## 3: P A Y A
## 4: P A Y A
## 5: P A Y A
## 6: P A Y A
## SURF_TYP SURF_TY2 TERRAIN COMP_DIR COUNTY FUNC_CLS MEDBARTY ST_FUNC
## 1: A M E 17 43 R1
## 2: A M E 17 43 R1
## 3: A M E 17 43 R1
## 4: A M E 17 43 R1
## 5: A M E 17 43 R1
## 6: A M E 17 43 R1
## RTE_NBR HPMS ROAD_INV SPD_LIMT BEGMP ENDMP LSHLDWID MEDWID
## 1: 2 5002172008272 002 55.07084 52.16 60.54 6.073922 0.8364139
## 2: 2 5002172008272 002 55.07084 52.16 60.54 6.073922 0.8364139
## 3: 2 5002172008272 002 55.07084 52.16 60.54 6.073922 0.8364139
## 4: 2 5002172008272 002 55.07084 52.16 60.54 6.073922 0.8364139
## 5: 2 5002172008272 002 55.07084 52.16 60.54 6.073922 0.8364139
## 6: 2 5002172008272 002 55.07084 52.16 60.54 6.073922 0.8364139
## NO_LANE1 NO_LANE2 NO_LANES RSHLDWID RSHL_WD2 SEG_LNG lanewid
## 1: 1.80421 1 3 4.867425 0.1539141 0.2689276 12.047
## 2: 1.80421 1 3 4.867425 0.1539141 0.2689276 12.047
## 3: 1.80421 1 3 4.867425 0.1539141 0.2689276 12.047
## 4: 1.80421 1 3 4.867425 0.1539141 0.2689276 12.047
## 5: 1.80421 1 3 4.867425 0.1539141 0.2689276 12.047
## 6: 1.80421 1 3 4.867425 0.1539141 0.2689276 12.047
## rdwy_wd1 rdwy_wd2 rdwy_wid AADT mvmt rodwycls ORN_FID_1 Total
## 1: 33.41593 0.489148 33.90508 5086.424 0.4998998 7 2638.708 16
## 2: 33.41593 0.489148 33.90508 5086.424 0.4998998 7 2638.708 16
## 3: 33.41593 0.489148 33.90508 5086.424 0.4998998 7 2638.708 16
## 4: 33.41593 0.489148 33.90508 5086.424 0.4998998 7 2638.708 16
## 5: 33.41593 0.489148 33.90508 5086.424 0.4998998 7 2638.708 16
## 6: 33.41593 0.489148 33.90508 5086.424 0.4998998 7 2638.708 16
## Fatal Injury PDO DAYMTH Crash spd_av spd_pv spd_ft date Month Day
## 1: 1 5 10 0101 0 NA NA NA 01012015 01 01
## 2: 1 5 10 0101 0 NA NA NA 01012015 01 01
## 3: 1 5 10 0101 0 NA NA NA 01012015 01 01
## 4: 1 5 10 0101 0 NA NA NA 01012015 01 01
## 5: 1 5 10 0101 0 NA NA NA 01012015 01 01
## 6: 1 5 10 0101 0 NA NA NA 01012015 01 01
## Year MonthDay
## 1: 2015 01_01
## 2: 2015 01_01
## 3: 2015 01_01
## 4: 2015 01_01
## 5: 2015 01_01
## 6: 2015 01_01
day1<- df_RMD[,-c(1)] %>% group_by(MonthDay) %>% summarize(Speed_All_Mean=mean(spd_av, na.rm=TRUE))
day1
## # A tibble: 365 x 2
## MonthDay Speed_All_Mean
## <chr> <dbl>
## 1 01_01 58.7
## 2 01_02 56.4
## 3 01_03 57.6
## 4 01_04 54.7
## 5 01_05 55.9
## 6 01_06 56.7
## 7 01_07 56.5
## 8 01_08 56.8
## 9 01_09 56.9
## 10 01_10 57.6
## # ... with 355 more rows
# Step 2: Examine Data
speed_clean <- tsclean(ts(day1$Speed_All_Mean))
plot.ts(speed_clean)
# ggplot() + geom_line(data = Q1, aes(x = TimeStamp, y = speed_clean)) + ylab('Cleaned Speed Records')
day1$cnt_ma = ma(speed_clean, order=7) # using the clean count with no outliers
day1$cnt_ma30 = ma(speed_clean, order=30)
# Step 3: Decompose Your Data
count_ma = ts(na.omit(speed_clean), frequency=30)
decomp = stl(count_ma, s.window="periodic")
deseasonal_cnt <- seasadj(decomp)
plot(decomp)
# Step 4: Stationarity
# statinary test
adf.test(count_ma, alternative = "stationary")
##
## Augmented Dickey-Fuller Test
##
## data: count_ma
## Dickey-Fuller = -2.8345, Lag order = 7, p-value = 0.2247
## alternative hypothesis: stationary
adf.test(deseasonal_cnt, alternative = "stationary")
##
## Augmented Dickey-Fuller Test
##
## data: deseasonal_cnt
## Dickey-Fuller = -2.6891, Lag order = 7, p-value = 0.2861
## alternative hypothesis: stationary
d1 = diff(deseasonal_cnt)
adf.test(d1, alternative = "stationary")
## Warning in adf.test(d1, alternative = "stationary"): p-value smaller than
## printed p-value
##
## Augmented Dickey-Fuller Test
##
## data: d1
## Dickey-Fuller = -10.33, Lag order = 7, p-value = 0.01
## alternative hypothesis: stationary
# Step 5: Autocorrelations and Choosing Model Order
# check ACF and PACF
acf2(count_ma)
## ACF PACF
## [1,] 0.49 0.49
## [2,] 0.13 -0.15
## [3,] 0.07 0.10
## [4,] 0.06 0.00
## [5,] 0.13 0.13
## [6,] 0.39 0.37
## [7,] 0.66 0.48
## [8,] 0.36 -0.16
## [9,] 0.06 -0.07
## [10,] 0.03 0.00
## [11,] 0.00 -0.10
## [12,] 0.04 -0.08
## [13,] 0.33 0.18
## [14,] 0.62 0.30
## [15,] 0.35 -0.06
## [16,] 0.05 -0.01
## [17,] -0.03 -0.11
## [18,] -0.01 0.06
## [19,] 0.05 -0.01
## [20,] 0.31 0.03
## [21,] 0.58 0.15
## [22,] 0.26 -0.18
## [23,] 0.00 0.01
## [24,] -0.03 0.01
## [25,] -0.06 -0.05
## [26,] -0.01 -0.04
## [27,] 0.26 0.08
## [28,] 0.54 0.13
## [29,] 0.27 -0.03
## [30,] -0.04 -0.09
## [31,] -0.11 -0.09
## [32,] -0.10 0.00
## [33,] -0.03 0.02
## [34,] 0.25 0.02
## [35,] 0.49 0.06
## [36,] 0.22 -0.03
## [37,] -0.06 -0.01
## [38,] -0.12 -0.04
## [39,] -0.11 0.00
## [40,] -0.06 -0.05
## [41,] 0.20 0.04
## [42,] 0.44 0.01
## [43,] 0.19 -0.04
## [44,] -0.08 -0.02
## [45,] -0.13 0.04
## [46,] -0.12 0.05
## [47,] -0.07 -0.03
## [48,] 0.19 0.00
## [49,] 0.43 0.05
## [50,] 0.18 -0.02
## [51,] -0.09 0.02
## [52,] -0.14 -0.05
## [53,] -0.12 0.02
## [54,] -0.08 -0.02
## [55,] 0.15 -0.05
## [56,] 0.41 0.08
## [57,] 0.15 -0.07
## [58,] -0.12 0.00
## [59,] -0.17 0.00
## [60,] -0.16 -0.02
## [61,] -0.10 -0.03
## [62,] 0.13 0.00
## [63,] 0.38 0.03
## [64,] 0.13 -0.05
## [65,] -0.15 -0.02
## [66,] -0.18 0.00
## [67,] -0.18 -0.04
## [68,] -0.13 0.00
## [69,] 0.12 0.01
## [70,] 0.35 0.01
## [71,] 0.10 -0.02
## [72,] -0.14 0.05
## [73,] -0.20 -0.04
## [74,] -0.20 -0.04
## [75,] -0.13 0.02
## [76,] 0.11 0.02
## [77,] 0.33 -0.02
## [78,] 0.09 -0.02
## [79,] -0.15 0.03
## [80,] -0.19 0.00
## [81,] -0.20 -0.02
## [82,] -0.14 -0.04
## [83,] 0.08 -0.01
## [84,] 0.29 -0.01
## [85,] 0.07 0.00
## [86,] -0.17 0.00
## [87,] -0.20 0.00
## [88,] -0.21 0.01
## [89,] -0.17 -0.05
## [90,] 0.06 -0.01
## [91,] 0.28 0.00
## [92,] 0.05 -0.01
## [93,] -0.17 0.03
## [94,] -0.22 -0.02
## [95,] -0.20 0.04
## [96,] -0.16 -0.03
## [97,] 0.07 0.03
## [98,] 0.26 -0.03
## [99,] 0.05 -0.01
## [100,] -0.17 -0.01
## [101,] -0.22 -0.02
## [102,] -0.21 0.03
## [103,] -0.16 -0.01
## [104,] 0.05 -0.02
## [105,] 0.25 0.00
## [106,] 0.04 0.00
## [107,] -0.18 -0.01
## [108,] -0.21 0.03
## [109,] -0.19 0.01
## [110,] -0.14 0.03
## [111,] 0.04 -0.06
## [112,] 0.23 -0.01
## [113,] 0.04 -0.01
## [114,] -0.17 -0.01
## [115,] -0.22 -0.02
## [116,] -0.20 -0.01
## [117,] -0.15 0.02
## [118,] 0.03 -0.02
## [119,] 0.22 -0.01
## [120,] 0.03 0.00
## ACF PACF
## [1,] 0.50 0.50
## [2,] 0.13 -0.16
## [3,] 0.08 0.11
## [4,] 0.07 -0.01
## [5,] 0.13 0.13
## [6,] 0.40 0.38
## [7,] 0.68 0.49
## [8,] 0.37 -0.17
## [9,] 0.06 -0.09
## [10,] 0.03 0.01
## [11,] 0.01 -0.10
## [12,] 0.04 -0.07
## [13,] 0.35 0.19
## [14,] 0.63 0.29
## [15,] 0.35 -0.08
## [16,] 0.05 -0.01
## [17,] -0.03 -0.12
## [18,] -0.01 0.05
## [19,] 0.05 0.00
## [20,] 0.31 0.02
## [21,] 0.59 0.16
## [22,] 0.27 -0.18
## [23,] -0.01 0.00
## [24,] -0.03 0.01
## [25,] -0.05 -0.04
## [26,] 0.00 -0.04
## [27,] 0.27 0.08
## [28,] 0.55 0.13
## [29,] 0.28 -0.04
## [30,] -0.05 -0.10
## [31,] -0.10 -0.09
## [32,] -0.10 -0.01
## [33,] -0.03 0.02
## [34,] 0.25 0.01
## [35,] 0.49 0.06
## [36,] 0.22 -0.01
## [37,] -0.06 0.01
## [38,] -0.12 -0.04
## [39,] -0.11 0.00
## [40,] -0.06 -0.05
## [41,] 0.20 0.04
## [42,] 0.45 0.02
## [43,] 0.20 -0.02
## [44,] -0.08 -0.01
## [45,] -0.13 0.04
## [46,] -0.12 0.03
## [47,] -0.07 -0.04
## [48,] 0.19 0.00
## [49,] 0.44 0.05
## [50,] 0.18 -0.03
## [51,] -0.09 0.04
## [52,] -0.14 -0.05
## [53,] -0.13 0.01
## [54,] -0.08 -0.03
## [55,] 0.15 -0.04
## [56,] 0.41 0.09
## [57,] 0.16 -0.07
## [58,] -0.12 0.01
## [59,] -0.16 -0.01
## [60,] -0.16 -0.03
## [61,] -0.11 -0.04
## [62,] 0.13 0.00
## [63,] 0.38 0.04
## [64,] 0.13 -0.06
## [65,] -0.16 -0.02
## [66,] -0.18 0.01
## [67,] -0.18 -0.02
## [68,] -0.13 0.00
## [69,] 0.11 0.01
## [70,] 0.35 0.01
## [71,] 0.11 -0.02
## [72,] -0.14 0.06
## [73,] -0.19 -0.03
## [74,] -0.19 -0.03
## [75,] -0.13 0.02
## [76,] 0.10 0.01
## [77,] 0.33 -0.04
## [78,] 0.09 -0.01
## [79,] -0.15 0.04
## [80,] -0.19 -0.01
## [81,] -0.20 0.01
## [82,] -0.15 -0.04
## [83,] 0.07 -0.01
## [84,] 0.29 -0.02
## [85,] 0.08 0.00
## [86,] -0.17 0.00
## [87,] -0.20 0.00
## [88,] -0.21 0.01
## [89,] -0.18 -0.07
## [90,] 0.05 -0.01
## [91,] 0.27 0.00
## [92,] 0.05 -0.02
## [93,] -0.17 0.04
## [94,] -0.22 -0.03
## [95,] -0.20 0.05
## [96,] -0.16 -0.03
## [97,] 0.06 0.05
## [98,] 0.26 -0.05
## [99,] 0.05 -0.01
## [100,] -0.17 -0.01
## [101,] -0.22 -0.02
## [102,] -0.20 0.05
## [103,] -0.15 0.01
## [104,] 0.05 -0.02
## [105,] 0.25 -0.01
## [106,] 0.04 -0.01
## [107,] -0.18 -0.02
## [108,] -0.21 0.04
## [109,] -0.19 0.02
## [110,] -0.15 0.01
## [111,] 0.04 -0.03
## [112,] 0.24 -0.02
## [113,] 0.04 -0.01
## [114,] -0.18 -0.02
## [115,] -0.21 -0.01
## [116,] -0.20 -0.01
## [117,] -0.15 0.03
## [118,] 0.03 -0.01
## [119,] 0.22 -0.02
## [120,] 0.02 -0.01
Seasonility Not in Consideration
# Step 6: Fitting an ARIMA model
auto.arima(deseasonal_cnt, seasonal=FALSE)
## Series: deseasonal_cnt
## ARIMA(1,1,5)
##
## Coefficients:
## ar1 ma1 ma2 ma3 ma4 ma5
## -0.5514 0.015 -0.9499 -0.4531 0.1934 0.3637
## s.e. 0.1155 0.114 0.0689 0.1084 0.0468 0.0679
##
## sigma^2 estimated as 0.4451: log likelihood=-367.57
## AIC=749.14 AICc=749.45 BIC=776.42
# Step 7: Evaluate and Iterate
# (try different model)
fit<-auto.arima(deseasonal_cnt, seasonal=FALSE)
tsdisplay(residuals(fit), lag.max=45, main='Model Residuals [Seasonality not considered]')
# step 8 forcast
fcast <- forecast(fit, h=30)
plot(fcast)
Seasonility in Consideration
# Step 6: Fitting an ARIMA model
auto.arima(deseasonal_cnt, seasonal=TRUE)
## Series: deseasonal_cnt
## ARIMA(1,1,3)(0,0,2)[30]
##
## Coefficients:
## ar1 ma1 ma2 ma3 sma1 sma2
## -0.7027 0.1690 -0.7765 -0.2002 -0.3791 -0.2691
## s.e. 0.1942 0.2113 0.0833 0.1163 0.0641 0.0753
##
## sigma^2 estimated as 0.4219: log likelihood=-364.12
## AIC=742.24 AICc=742.56 BIC=769.52
# Step 7: Evaluate and Iterate
# (try different model)
fit<-auto.arima(deseasonal_cnt, seasonal=TRUE)
tsdisplay(residuals(fit), lag.max=45, main='Model Residuals [Seasonality considered]')
# step 8 forcast
fcast <- forecast(fit, h=30)
plot(fcast)
