140SL MIDTERM
JIASHU MIAO 804786709
2019/2/6
## 1.
# rm(list=ls())
pkg <- c("readr","readxl","dplyr","stringr","ggplot2","tidyr")
pkgload <- lapply(pkg, require, character.only = TRUE)## Loading required package: readr## Loading required package: readxl## Loading required package: dplyr##
## Attaching package: 'dplyr'## The following objects are masked from 'package:stats':
##
## filter, lag## The following objects are masked from 'package:base':
##
## intersect, setdiff, setequal, union## Loading required package: stringr## Loading required package: ggplot2## Loading required package: tidyr- Provide a demographic profile for both test components (AnionGap and SODIUM).
data1 <- read_excel("/Users/MichaelMiao/UCLA/uclatextbook/140sl/140\ review\ from\ Wilbur/demographicData4pm.xls")## readxl works best with a newer version of the tibble package.
## You currently have tibble v1.4.2.
## Falling back to column name repair from tibble <= v1.4.2.
## Message displays once per session.load("/Users/MichaelMiao/UCLA/uclatextbook/140sl/140\ review\ from\ Wilbur/LAB4PM.RData")
data2 <- LAB4PM
summary(data1)## STUDY_ID Gender Race
## Length:18721 Length:18721 Length:18721
## Class :character Class :character Class :character
## Mode :character Mode :character Mode :character
## Ethnicity
## Length:18721
## Class :character
## Mode :charactersummary(data2)## STUDY_ID AnionGap Age Year
## Length:14161 Min. : 4.00 Min. :30.00 Min. :2016
## Class :character 1st Qu.:12.00 1st Qu.:32.00 1st Qu.:2016
## Mode :character Median :14.00 Median :34.00 Median :2016
## Mean :13.77 Mean :34.45 Mean :2016
## 3rd Qu.:15.00 3rd Qu.:37.00 3rd Qu.:2017
## Max. :35.00 Max. :40.00 Max. :2017
## Inpatient_Outpatient SODIUM
## Length:14161 Min. :123.0
## Class :character 1st Qu.:139.0
## Mode :character Median :140.0
## Mean :140.1
## 3rd Qu.:142.0
## Max. :151.0names(data1)## [1] "STUDY_ID" "Gender" "Race" "Ethnicity"names(data2)## [1] "STUDY_ID" "AnionGap" "Age"
## [4] "Year" "Inpatient_Outpatient" "SODIUM"datajoin <- inner_join(data1,data2,by = "STUDY_ID")
#View(datajoin)
datajoind <- distinct(datajoin)
attach(datajoin)
summary(datajoin)## STUDY_ID Gender Race
## Length:14161 Length:14161 Length:14161
## Class :character Class :character Class :character
## Mode :character Mode :character Mode :character
##
##
##
## Ethnicity AnionGap Age Year
## Length:14161 Min. : 4.00 Min. :30.00 Min. :2016
## Class :character 1st Qu.:12.00 1st Qu.:32.00 1st Qu.:2016
## Mode :character Median :14.00 Median :34.00 Median :2016
## Mean :13.77 Mean :34.45 Mean :2016
## 3rd Qu.:15.00 3rd Qu.:37.00 3rd Qu.:2017
## Max. :35.00 Max. :40.00 Max. :2017
## Inpatient_Outpatient SODIUM
## Length:14161 Min. :123.0
## Class :character 1st Qu.:139.0
## Mode :character Median :140.0
## Mean :140.1
## 3rd Qu.:142.0
## Max. :151.0tapply(datajoin$AnionGap,datajoin$Race,mean)## American Indian or Alaska Native
## 14.23913
## Asian
## 13.83436
## Black or African American
## 13.15815
## Multiple Races
## 13.61905
## NA
## 13.13636
## Native Hawaiian or Other Pacific Islander
## 14.45946
## Other
## 13.80646
## Patient Refused
## 14.21692
## Unknown
## 13.92087
## White or Caucasian
## 13.65992tapply(datajoin$AnionGap,datajoin$Gender,mean)## Female Male Unknown
## 13.67799 13.87348 16.00000tapply(datajoin$AnionGap,datajoin$Ethnicity,mean)## Cuban Hispanic or Latino
## 16.00000 13.59548
## Hispanic/Spanish origin Other Mexican, Mexican American, Chicano/a
## 14.01170 13.00000
## Not Hispanic or Latino Patient Refused
## 13.69819 14.19592
## Puerto Rican Unknown
## 12.50000 13.93228tapply(datajoin$SODIUM,datajoin$Race,mean)## American Indian or Alaska Native
## 139.3261
## Asian
## 140.1168
## Black or African American
## 140.0032
## Multiple Races
## 139.7738
## NA
## 139.9545
## Native Hawaiian or Other Pacific Islander
## 139.7297
## Other
## 140.1326
## Patient Refused
## 140.2988
## Unknown
## 140.2171
## White or Caucasian
## 140.1332tapply(datajoin$SODIUM,datajoin$Gender,mean)## Female Male Unknown
## 139.7425 140.6421 143.0000tapply(datajoin$SODIUM,datajoin$Ethnicity,mean)## Cuban Hispanic or Latino
## 141.2500 139.9774
## Hispanic/Spanish origin Other Mexican, Mexican American, Chicano/a
## 140.2047 139.7419
## Not Hispanic or Latino Patient Refused
## 140.1228 140.3020
## Puerto Rican Unknown
## 140.7500 140.2659prop.table(table(datajoin$Gender))##
## Female Male Unknown
## 5.528564e-01 4.470729e-01 7.061648e-05prop.table(table(datajoin$Race))##
## American Indian or Alaska Native
## 0.003248358
## Asian
## 0.102746981
## Black or African American
## 0.044205918
## Multiple Races
## 0.005931784
## NA
## 0.003107125
## Native Hawaiian or Other Pacific Islander
## 0.002612810
## Other
## 0.170397571
## Patient Refused
## 0.078454911
## Unknown
## 0.149918791
## White or Caucasian
## 0.439375750table(Gender)## Gender
## Female Male Unknown
## 7829 6331 1table(Race)## Race
## American Indian or Alaska Native
## 46
## Asian
## 1455
## Black or African American
## 626
## Multiple Races
## 84
## NA
## 44
## Native Hawaiian or Other Pacific Islander
## 37
## Other
## 2413
## Patient Refused
## 1111
## Unknown
## 2123
## White or Caucasian
## 6222table(Ethnicity)## Ethnicity
## Cuban Hispanic or Latino
## 4 1372
## Hispanic/Spanish origin Other Mexican, Mexican American, Chicano/a
## 171 124
## Not Hispanic or Latino Patient Refused
## 9042 1225
## Puerto Rican Unknown
## 8 2215datajoinnew <- datajoin %>% mutate(.,Noinfo = replace(Race, Race=="Patient Refused","Unknown"))
table(datajoinnew$Noinfo)##
## American Indian or Alaska Native
## 46
## Asian
## 1455
## Black or African American
## 626
## Multiple Races
## 84
## NA
## 44
## Native Hawaiian or Other Pacific Islander
## 37
## Other
## 2413
## Unknown
## 3234
## White or Caucasian
## 6222datajoinnew <- datajoinnew %>% select_at(.,vars(-c(Race)))
datajoinnew <- rename(datajoinnew, Race = Noinfo)
dim(datajoinnew)## [1] 14161 9datajoinnew <- datajoinnew %>% mutate(., NO = replace(Ethnicity, Ethnicity == "Patient Refused","Unknown"))
datajoinnew <- datajoinnew %>% select_at(.,vars(-c(Ethnicity)))
dim(datajoinnew)## [1] 14161 9datajoinnew <- rename(datajoinnew, Ethnicity = NO)- There are about 55% females and 44% males enroll in this test. There are many unknows and I would like to make the patient refused as unknown for both race and ethicnicity. The two component seems to have no specific pattern but they distribute well.
str(datajoinnew)## Classes 'tbl_df', 'tbl' and 'data.frame': 14161 obs. of 9 variables:
## $ STUDY_ID : chr "000EA425FFF3622052E55966D027AEC7" "0013E732CFD6284BBBDC47C8B3D6132A" "0014D2A96ECDA4C63B10E375F26DB63A" "00162780B66CD4C246AE00EA683D2CE9" ...
## $ Gender : chr "Female" "Female" "Female" "Female" ...
## $ AnionGap : num 13 15 13 15 14 15 16 14 12 11 ...
## $ Age : int 33 35 38 39 33 33 32 31 35 32 ...
## $ Year : int 2016 2016 2016 2016 2016 2017 2016 2017 2017 2016 ...
## $ Inpatient_Outpatient: chr "IP" "OP" "OP" "OP" ...
## $ SODIUM : num 141 140 141 143 141 142 141 144 139 136 ...
## $ Race : chr "Other" "Other" "White or Caucasian" "White or Caucasian" ...
## $ Ethnicity : chr "Not Hispanic or Latino" "Unknown" "Not Hispanic or Latino" "Not Hispanic or Latino" ...par(mfrow=c(1,2))
ggplot(data = datajoinnew,aes(factor(Year),AnionGap))+geom_boxplot(fill = c("yellow","blue"))
ggplot(data = datajoinnew,aes(factor(Year),SODIUM))+geom_boxplot(fill = c("yellow","blue"))
attach(datajoinnew)## The following objects are masked from datajoin:
##
## Age, AnionGap, Ethnicity, Gender, Inpatient_Outpatient, Race,
## SODIUM, STUDY_ID, Yearplot(SODIUM,Year)
plot(AnionGap,Year)
plot(as.factor(Gender),AnionGap)
plot(as.factor(Gender),SODIUM)
plot(as.factor(Race),AnionGap)
plot(as.factor(Race),SODIUM)
plot(as.factor(Ethnicity),AnionGap)
plot(as.factor(Ethnicity),SODIUM)
attach(datajoinnew)## The following objects are masked from datajoinnew (pos = 3):
##
## Age, AnionGap, Ethnicity, Gender, Inpatient_Outpatient, Race,
## SODIUM, STUDY_ID, Year## The following objects are masked from datajoin:
##
## Age, AnionGap, Ethnicity, Gender, Inpatient_Outpatient, Race,
## SODIUM, STUDY_ID, Yearmodel1 <- glm(data = datajoinnew, SODIUM~ factor(Ethnicity)+ factor(Race) + factor(Gender))
summary(model1)##
## Call:
## glm(formula = SODIUM ~ factor(Ethnicity) + factor(Race) + factor(Gender),
## data = datajoinnew)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -17.2885 -1.6203 0.2097 1.3797 10.3797
##
## Coefficients:
## Estimate Std. Error
## (Intercept) 140.12018 1.23701
## factor(Ethnicity)Hispanic or Latino -1.19461 1.18805
## factor(Ethnicity)Hispanic/Spanish origin Other -0.99804 1.20011
## factor(Ethnicity)Mexican, Mexican American, Chicano/a -1.39230 1.20539
## factor(Ethnicity)Not Hispanic or Latino -1.06050 1.18649
## factor(Ethnicity)Puerto Rican -0.49213 1.45296
## factor(Ethnicity)Unknown -0.88782 1.18953
## factor(Race)Asian 0.73789 0.35554
## factor(Race)Black or African American 0.54781 0.36271
## factor(Race)Multiple Races 0.41494 0.43520
## factor(Race)NA 0.25916 0.50791
## factor(Race)Native Hawaiian or Other Pacific Islander 0.27560 0.52417
## factor(Race)Other 0.73058 0.35342
## factor(Race)Unknown 0.59755 0.36162
## factor(Race)White or Caucasian 0.66737 0.35121
## factor(Gender)Male 0.89330 0.04027
## factor(Gender)Unknown 3.34386 2.37335
## t value Pr(>|t|)
## (Intercept) 113.273 <2e-16 ***
## factor(Ethnicity)Hispanic or Latino -1.006 0.3147
## factor(Ethnicity)Hispanic/Spanish origin Other -0.832 0.4056
## factor(Ethnicity)Mexican, Mexican American, Chicano/a -1.155 0.2481
## factor(Ethnicity)Not Hispanic or Latino -0.894 0.3714
## factor(Ethnicity)Puerto Rican -0.339 0.7348
## factor(Ethnicity)Unknown -0.746 0.4555
## factor(Race)Asian 2.075 0.0380 *
## factor(Race)Black or African American 1.510 0.1310
## factor(Race)Multiple Races 0.953 0.3404
## factor(Race)NA 0.510 0.6099
## factor(Race)Native Hawaiian or Other Pacific Islander 0.526 0.5990
## factor(Race)Other 2.067 0.0387 *
## factor(Race)Unknown 1.652 0.0985 .
## factor(Race)White or Caucasian 1.900 0.0574 .
## factor(Gender)Male 22.185 <2e-16 ***
## factor(Gender)Unknown 1.409 0.1589
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 5.627735)
##
## Null deviance: 82581 on 14160 degrees of freedom
## Residual deviance: 79599 on 14144 degrees of freedom
## AIC: 64672
##
## Number of Fisher Scoring iterations: 2model2 <- glm(data = datajoinnew, AnionGap~ factor(Ethnicity)+ factor(Race) + factor(Gender))
summary(model2)##
## Call:
## glm(formula = AnionGap ~ factor(Ethnicity) + factor(Race) + factor(Gender),
## data = datajoinnew)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -9.9460 -1.7726 0.0458 1.4140 21.4140
##
## Coefficients:
## Estimate Std. Error
## (Intercept) 16.48070 1.32351
## factor(Ethnicity)Hispanic or Latino -2.46108 1.27112
## factor(Ethnicity)Hispanic/Spanish origin Other -2.05224 1.28403
## factor(Ethnicity)Mexican, Mexican American, Chicano/a -3.06385 1.28968
## factor(Ethnicity)Not Hispanic or Latino -2.26967 1.26945
## factor(Ethnicity)Puerto Rican -3.50846 1.55455
## factor(Ethnicity)Unknown -2.07179 1.27271
## factor(Race)Asian -0.45157 0.38040
## factor(Race)Black or African American -1.13579 0.38807
## factor(Race)Multiple Races -0.60796 0.46563
## factor(Race)NA -1.36555 0.54342
## factor(Race)Native Hawaiian or Other Pacific Islander 0.17267 0.56082
## factor(Race)Other -0.42089 0.37814
## factor(Race)Unknown -0.45472 0.38690
## factor(Race)White or Caucasian -0.62499 0.37577
## factor(Gender)Male 0.18653 0.04308
## factor(Gender)Unknown 2.40128 2.53931
## t value Pr(>|t|)
## (Intercept) 12.452 < 2e-16 ***
## factor(Ethnicity)Hispanic or Latino -1.936 0.05287 .
## factor(Ethnicity)Hispanic/Spanish origin Other -1.598 0.11000
## factor(Ethnicity)Mexican, Mexican American, Chicano/a -2.376 0.01753 *
## factor(Ethnicity)Not Hispanic or Latino -1.788 0.07381 .
## factor(Ethnicity)Puerto Rican -2.257 0.02403 *
## factor(Ethnicity)Unknown -1.628 0.10358
## factor(Race)Asian -1.187 0.23521
## factor(Race)Black or African American -2.927 0.00343 **
## factor(Race)Multiple Races -1.306 0.19168
## factor(Race)NA -2.513 0.01199 *
## factor(Race)Native Hawaiian or Other Pacific Islander 0.308 0.75818
## factor(Race)Other -1.113 0.26570
## factor(Race)Unknown -1.175 0.23991
## factor(Race)White or Caucasian -1.663 0.09629 .
## factor(Gender)Male 4.330 1.5e-05 ***
## factor(Gender)Unknown 0.946 0.34435
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 6.442289)
##
## Null deviance: 92020 on 14160 degrees of freedom
## Residual deviance: 91120 on 14144 degrees of freedom
## AIC: 66586
##
## Number of Fisher Scoring iterations: 2model3 <- glm(data = datajoinnew,SODIUM~AnionGap)
summary(model3)##
## Call:
## glm(formula = SODIUM ~ AnionGap, data = datajoinnew)
##
## Deviance Residuals:
## Min 1Q Median 3Q Max
## -16.9951 -1.4079 0.0049 1.5921 10.3964
##
## Coefficients:
## Estimate Std. Error t value Pr(>|t|)
## (Intercept) 137.45071 0.10905 1260.44 <2e-16 ***
## AnionGap 0.19572 0.00779 25.13 <2e-16 ***
## ---
## Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
##
## (Dispersion parameter for gaussian family taken to be 5.583425)
##
## Null deviance: 82581 on 14160 degrees of freedom
## Residual deviance: 79056 on 14159 degrees of freedom
## AIC: 64545
##
## Number of Fisher Scoring iterations: 2I think there is very little association of Sodium between gender(male) and race(asian) ; and little association of Anigap between gender(male) and race(black). The association is very tiny. There is association between two Sodium and Aniongap. The value goes up with year as well.