BST 210 Project
2022-10-08
Data Cleaning
data$AGE[data$AGE == 997] = NA
data$AGE[data$AGE == 999] = NA
data$AGE[data$AGE == 0] = NAdata$sexcat <- NA
data$sexcat[data$SEX == 1] = "male"
data$sexcat[data$SEX == 2] = "female"
data$sexcat <- as.factor(data$sexcat)data$sexoricat <- NA
data$sexoricat[data$SEXORIEN == 1] = "homosexual"
data$sexoricat[data$SEXORIEN == 2] = "heterosexual"
data$sexoricat[data$SEXORIEN == 3] = "bisexual"
data$sexoricat[data$SEXORIEN == 4] = "other"
data$sexoricat <- factor(data$sexoricat, levels = c("homosexual", "heterosexual", "bisexual", "other"))data$racecat <- NA
data$racecat[data$RACENEW == 100] = "white"
data$racecat[data$RACENEW == 200] = "black"
data$racecat[data$RACENEW == 300] = "aian"
data$racecat[data$RACENEW == 400] = "asian"
data$racecat[data$RACENEW > 400 & data$RACENEW < 900] = "other"
data$racecat <- as.factor(data$racecat)data$HOURSWRK[data$HOURSWRK > 95] = NA
data$HOURSWRK[data$HOURSWRK == 0] = NAdata$incomecat <- NA
data$incomecat[data$INCFAM07ON == 11] = "under $34,999"
data$incomecat[data$INCFAM07ON == 12] = "$35,000-$49,999"
data$incomecat[data$INCFAM07ON == 22] = "$50,000-$74,999"
data$incomecat[data$INCFAM07ON == 23] = "$75,000-$99,999"
data$incomecat[data$INCFAM07ON == 24] = "over $100,000"
data$incomecat <- factor(data$incomecat, levels = c("under $34,999", "$35,000-$49,999", "$50,000-$74,999", "$75,000-$99,999", "over $100,000"))data$healthcat <- NA
data$healthcat[data$HEALTH == 1] = "excellent"
data$healthcat[data$HEALTH == 2] = "very good"
data$healthcat[data$HEALTH == 3] = "good"
data$healthcat[data$HEALTH == 4] = "fair"
data$healthcat[data$HEALTH == 5] = "poor"
data$healthcat <- factor(data$healthcat, levels = c("excellent", "very good", "good", "fair", "poor"))data$insurcat <- NA
data$insurcat[data$HINOTCOVE == 1] = "no"
data$insurcat[data$HINOTCOVE == 2] = "yes"
data$insurcat <- as.factor(data$insurcat)data$MOD10DMIN[data$MOD10DMIN %in% c(0, 996, 997, 998, 999)] = NAdata$MOD10FWK[data$MOD10FWK == 0] = NA
data$MOD10FWK[data$MOD10FWK > 90] = NAdata$anxiety <- NA
data$anxiety[data$ANXIETYEV == 1] = "No"
data$anxiety[data$ANXIETYEV == 2] = "Yes"
data$anxiety <- as.factor(data$anxiety)data$depress <- NA
data$depress[data$DEPFREQ == 1] = "daily"
data$depress[data$DEPFREQ == 2] = "weekly"
data$depress[data$DEPFREQ == 3] = "monthly"
data$depress[data$DEPFREQ == 4] = "yearly"
data$depress[data$DEPFREQ == 5] = "never"
data$depress <- factor(data$depress, levels = c("daily", "weekly", "monthly", "yearly", "never"))data$esupport <- NA
data$esupport[data$EMOSUPPORT == 1] = "always"
data$esupport[data$EMOSUPPORT == 2] = "usually"
data$esupport[data$EMOSUPPORT == 3] = "sometimes"
data$esupport[data$EMOSUPPORT == 4] = "rarely"
data$esupport[data$EMOSUPPORT == 5] = "never"
data$esupport <- factor(data$esupport, levels = c("always", "usually", "sometimes", "rarely", "never"))data$covid <- NA
data$covid[data$CVDTESTRSLT == 1] = "No"
data$covid[data$CVDTESTRSLT == 2] = "Yes"
data$covid <- factor(data$covid, levels = c("Yes", "No"))data$sympt <- NA
data$sympt[data$CVDSYMP == 1] = "no symptoms"
data$sympt[data$CVDSYMP == 2] = "mild symptoms"
data$sympt[data$CVDSYMP == 3] = "moderate symptoms"
data$sympt[data$CVDSYMP == 4] = "severe symptoms"
data$sympt <- factor(data$sympt, levels = c("no symptoms", "mild symptoms", "moderate symptoms", "severe symptoms"))data %>%
select(31:41, AGE, HOURSWRK, MOD10DMIN, MOD10FWK, HISPETH, EDUC) %>%
summary()## sexcat sexoricat racecat incomecat
## female:19865 homosexual : 571 aian : 287 under $34,999 : 9136
## male :17487 heterosexual:29361 asian: 2090 $35,000-$49,999: 4641
## NA's : 6 bisexual : 407 black: 3916 $50,000-$74,999: 6699
## other : 149 other: 1057 $75,000-$99,999: 4855
## NA's : 6870 white:28047 over $100,000 :12027
## NA's : 1961
##
## healthcat insurcat anxiety depress esupport
## excellent:11071 no :34595 No :26617 daily : 1284 always : 8573
## very good:12269 yes : 2682 Yes : 4900 weekly : 2013 usually : 4542
## good : 9342 NA's: 81 NA's: 5841 monthly: 2459 sometimes: 2226
## fair : 3576 yearly : 9596 rarely : 747
## poor : 1080 never :20003 never : 960
## NA's : 20 NA's : 2003 NA's :20310
##
## covid sympt AGE HOURSWRK
## Yes : 499 no symptoms : 99 Min. : 1 Min. : 1.0
## No : 4925 mild symptoms : 224 1st Qu.:30 1st Qu.:35.0
## NA's:31934 moderate symptoms: 238 Median :49 Median :40.0
## severe symptoms : 132 Mean :47 Mean :39.4
## NA's :36665 3rd Qu.:66 3rd Qu.:45.0
## Max. :85 Max. :95.0
## NA's :374 NA's :20223
## MOD10DMIN MOD10FWK HISPETH EDUC
## Min. : 1.00 Min. : 1.000 Min. :10.00 Min. : 0.0
## 1st Qu.: 30.00 1st Qu.: 2.000 1st Qu.:10.00 1st Qu.:201.0
## Median : 45.00 Median : 4.000 Median :10.00 Median :301.0
## Mean : 56.26 Mean : 4.485 Mean :13.84 Mean :270.6
## 3rd Qu.: 60.00 3rd Qu.: 7.000 3rd Qu.:10.00 3rd Qu.:400.0
## Max. :720.00 Max. :28.000 Max. :93.00 Max. :999.0
## NA's :15376 NA's :16372
4 numeric variables: age, total hours worked last week usually, duration of moderate activity 10+ minutues, frequency of moderate activity 10+ minutes
2 un-cleaned variables: hispanic ethnicity, educational attainment (we have not found an efficient manner of collapsing ethnicity and education levels)
Our outcome variables are inbalanced (even after dropping missing values) so we might need to downsample or upsample.
Data Visualization
data<- drop_na(data) We are not considering missing data for now.
data %>%
ggplot(aes(x=covid, fill=anxiety)) +
geom_bar(position="fill")
data %>%
ggplot(aes(x=covid, fill=depress)) +
geom_bar(position="fill")
Covid seems to be negatively associated with anxiety and depression.
data %>%
ggplot(aes(x=sexcat, fill=anxiety)) +
geom_bar(position="fill")
data %>%
ggplot(aes(x=sexcat, fill=depress)) +
geom_bar(position="fill")
Female seems to be positively associated with anxiety and depression.
data %>%
ggplot(aes(x=sexoricat, fill=anxiety)) +
geom_bar(position="fill")
data %>%
ggplot(aes(x=sexoricat, fill=depress)) +
geom_bar(position="fill")
Bisexual seems to be at higher risk of anxiety and depression, followed by homosexual and lastly heterosexual.
data %>%
ggplot(aes(x=racecat, fill=anxiety)) +
geom_bar(position="fill")
data %>%
ggplot(aes(x=racecat, fill=depress)) +
geom_bar(position="fill")
White and other seem to be at higher risk of anxiety and depression.
data %>%
ggplot(aes(x=incomecat, fill=anxiety)) +
geom_bar(position="fill")
data %>%
ggplot(aes(x=incomecat, fill=depress)) +
geom_bar(position="fill")
Higher income seems to be a protective factor.
data %>%
ggplot(aes(x=healthcat, fill=anxiety)) +
geom_bar(position="fill")
data %>%
ggplot(aes(x=healthcat, fill=depress)) +
geom_bar(position="fill")
Poor health status is highly associated with anxiety and especially depression.
data %>%
ggplot(aes(x=insurcat, fill=anxiety)) +
geom_bar(position="fill")
data %>%
ggplot(aes(x=insurcat, fill=depress)) +
geom_bar(position="fill")
Insurance seems to be a protective factor.
data %>%
ggplot(aes(x=esupport, fill=anxiety)) +
geom_bar(position="fill")
data %>%
ggplot(aes(x=esupport, fill=depress)) +
geom_bar(position="fill")
Participants who rarely or sometimes receive emotional support seem to be under higher risk of anxiety and depression.
data %>%
ggplot(aes(x=sympt, fill=anxiety)) +
geom_bar(position="fill")
data %>%
ggplot(aes(x=sympt, fill=depress)) +
geom_bar(position="fill")
Severity of covid symptoms seems to be positively associated with anxiety and depression.
data %>%
ggplot(aes(x=anxiety,y=AGE)) +
geom_boxplot()
data %>%
ggplot(aes(x=depress,y=AGE)) +
geom_boxplot()
Participants with anxiety are on average younger, while participants who feel depressed daily are older between 40 and 60.
data %>%
ggplot(aes(x=anxiety,y=HOURSWRK)) +
geom_boxplot()
data %>%
ggplot(aes(x=depress,y=HOURSWRK)) +
geom_boxplot()
Participants with anxiety or depression seem to work fewer hours.
data %>%
ggplot(aes(x=anxiety,y=MOD10DMIN)) +
geom_boxplot()
data %>%
ggplot(aes(x=depress,y=MOD10DMIN)) +
geom_boxplot()
Exercise seems to be a protective factor.
data %>%
ggplot(aes(x=anxiety,y=MOD10FWK)) +
geom_boxplot()
data %>%
ggplot(aes(x=depress,y=MOD10FWK)) +
geom_boxplot()