commute-health-study/CommCode.R at main · yoid2000/commute-health-study · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
#B. Lesko�ek, 2024-04-25.
#for B. Francis, synDiffix
#code used for paper https://link.springer.com/content/pdf/10.1186/s12889-021-10326-6.pdf

# Total columns: "VO2max","CommToSch","CommHome","gender","age","MVPAsqrt","DistLogToHome","DistLogToSch","DistFromHome","DistFromSchool"

runComputation <- function(commDataPath2Sch, commDataPath2Home, dataPath, plotPath) {
    library(ggplot2)
    library(cowplot)

    dd2s <- read.csv(commDataPath2Sch)
    # If dd2s$DistFromHome is in dd2s, the convert it to log
    if ("DistFromHome" %in% colnames(dd2s)) {
        print("Do the transformation for DistFromHome")
        dd2s$DistLogToSch <- log2(dd2s$DistFromHome)
    }
    dd2s$CommToSch = factor(dd2s$CommToSch, levels=c('car', 'public', 'wheels', 'walk'))
    dd2s$gender = factor(dd2s$gender, levels=c('male', 'female'))
    dd2s <- within(dd2s, gender <- relevel(gender, ref = 2))
    dd2s <- within(dd2s, CommToSch <- relevel(CommToSch, ref = 4))

    dd2h <- read.csv(commDataPath2Home)
    if ("DistFromSchool" %in% colnames(dd2h)) {
        print("Do the transformation for DistFromSchool")
        dd2h$DistLogToHome <- log2(dd2h$DistFromSchool)
    }
    dd2h$CommHome = factor(dd2h$CommHome, levels=c('car', 'public', 'wheels', 'walk'))
    dd2h$gender = factor(dd2h$gender, levels=c('male', 'female'))
    dd2h <- within(dd2h, gender <- relevel(gender, ref = 2))
    dd2h <- within(dd2h, CommHome <- relevel(CommHome, ref = 4))

    ### from Home 2 Sch
    m=lm(VO2max ~ CommToSch*gender + age + MVPAsqrt + DistLogToSch:CommToSch , data=dd2s)
    summary(m)
    summ <- summary(m)
    conf_int <- confint(m, level = 0.95)
    df2s <- as.data.frame(conf_int)
    df2s <- na.omit(df2s)
    if (nrow(df2s) == nrow(summ$coefficients)) {
        # If they match, proceed to assign
        df2s$Estimate <- summ$coefficients[, "Estimate"]
        df2s$Std.Error <- summ$coefficients[, "Std. Error"]
        df2s$t.value <- summ$coefficients[, "t value"]
        df2s$Pr.t. <- summ$coefficients[, "Pr(>|t|)"]
      } else {
        # If they don't match, print a warning or handle the discrepancy
        warning("Home 2 Sch Mismatch in number of coefficients and confidence intervals. Check model specification.")
        # Output the contents of df2s and summ$coefficients
        print("Contents of df2s (Confidence Intervals):")
        print(df2s)
        print("Contents of summ$coefficients:")
        print(summ$coefficients)
        # Optional: Handle the discrepancy, e.g., by aligning the rows or investigating the cause
      }
    library(jsonlite)
    new <- data.frame(gender=c(rep(1,4), rep(2,4)), CommToSch=0:3, age=mean(dd2s$age), MVPAsqrt=mean(dd2s$MVPAsqrt),
                    DistLogToSch=aggregate(dd2s[, "DistLogToSch"], list(dd2s$CommToSch), median)$x)
    new$gender= factor(new$gender);  levels(new$gender)=levels(dd2s$gender)
    new$CommToSch= factor(new$CommToSch);  levels(new$CommToSch)=levels(dd2s$CommToSch)
    new$CommToSch = factor(new$CommToSch,levels(dd2s$CommToSch)[c(1,2,4,3)])
    Ypred= predict.lm(m, new)
    #zzzz pred.w.plim <- predict(m, new, interval = "prediction")
    pred.w.clim <- predict(m, new, interval = "confidence")
    pd <- position_dodge(0.3) # move them .05 to the left and right
    q=cbind(new, pred.w.clim)
    q$CommToSch <- factor(q$CommToSch, levels = c("car", "public", "walk", "wheels"))
    q$gender <- factor(q$gender, levels = c("male", "female"))
    p1=ggplot(q, aes(x=CommToSch, y=fit, colour=gender)) +
        geom_errorbar(aes(ymin=lwr, ymax=upr), width=.1, position=pd) +
        geom_line(position=pd) +
        geom_point(position=pd) +
        ylim(39,57) +
        xlab("Commuting mode (from home to school)") +
        ylab("VO2max (predicted at mode median distance)") +
        theme(legend.position = "none")

    df_conf_2s <- data.frame(
    CommToSch = q$CommToSch,
    gender = q$gender,
    fit = q$fit,
    lwr = q$lwr,
    upr = q$upr
    )

    ### from Sch 2 Home
    m=lm(VO2max ~ CommHome*gender + age + MVPAsqrt + DistLogToHome:CommHome , data=dd2h)
    summary(m)
    summ <- summary(m)
    conf_int <- confint(m, level = 0.95)
    df2h <- as.data.frame(conf_int)
    df2h <- na.omit(df2h)
    if (nrow(df2h) == nrow(summ$coefficients)) {
        # If they match, proceed to assign
        df2h$Estimate <- summ$coefficients[, "Estimate"]
        df2h$Std.Error <- summ$coefficients[, "Std. Error"]
        df2h$t.value <- summ$coefficients[, "t value"]
        df2h$Pr.t. <- summ$coefficients[, "Pr(>|t|)"]
      } else {
        # If they don't match, print a warning or handle the discrepancy
        warning("Sch 2 Home Mismatch in number of coefficients and confidence intervals. Check model specification.")
        # Output the contents of df2h and summ$coefficients
        print("Contents of df2h (Confidence Intervals):")
        print(df2h)
        print("Contents of summ$coefficients:")
        print(summ$coefficients)
        # Optional: Handle the discrepancy, e.g., by aligning the rows or investigating the cause
      }
    new= data.frame(gender=c(rep(1,4), rep(2,4)), CommHome=0:3, age=mean(dd2h$age), MVPAsqrt=mean(dd2h$MVPAsqrt), DistLogToHome=aggregate(dd2h[, "DistLogToHome"], list(dd2h$CommHome), median)$x)
    new$gender= factor(new$gender);  levels(new$gender)=levels(dd2h$gender)
    new$CommHome= factor(new$CommHome);  levels(new$CommHome)=levels(dd2h$CommHome)
    new$CommHome = factor(new$CommHome,levels(dd2h$CommHome)[c(1,2,4,3)])
    Ypred= predict.lm(m, new)
    #zzzz pred.w.plim <- predict(m, new, interval = "prediction")
    pred.w.clim <- predict(m, new, interval = "confidence")
    pd <- position_dodge(0.3) # move them .05 to the left and right
    q=cbind(new, pred.w.clim)
    q$CommHome <- factor(q$CommHome, levels = c("car", "public", "walk", "wheels"))
    q$gender <- factor(q$gender, levels = c("male", "female"))
    p2=ggplot(q, aes(x=CommHome, y=fit, colour=gender)) +
        geom_errorbar(aes(ymin=lwr, ymax=upr), width=.1, position=pd) +
        geom_line(position=pd) +
        geom_point(position=pd) +
        ylim(39,57) +
        xlab("Commuting mode (from school to home)") +
        ylab(NULL) +
        guides(colour = guide_legend(title = NULL))  # Remove the legend title
    # Extract the legend from p2
    legend <- get_legend(p2)
    # Remove the legend from p2
    p2 <- p2 + theme(legend.position = "none")

    # Combine the plots and the legend
    combined_plot <- plot_grid(
        p1, p2, legend,
        ncol = 3,
        rel_widths = c(1, 1, 0.2)  # Adjust the relative widths as needed
    )
    # Save the combined plot
    png(plotPath, width = 20, height = 10, units = "cm", res = 600)
    print(combined_plot)
    dev.off()

    #library(gridExtra)
    #grid.arrange(p1, p2, ncol=2)
    #dev.off()

    df_conf_2h <- data.frame(
    CommHome = q$CommHome,
    gender = q$gender,
    fit = q$fit,
    lwr = q$lwr,
    upr = q$upr
    )

    # COmbine the stored data into one dataframe and save to json
    df_all <- list(
    coef_2s = df2s,
    conf_2s = df_conf_2s,
    coef_2h = df2h,
    conf_2h = df_conf_2h
    )
    write_json(df_all, path = dataPath, pretty=TRUE)
}

options(error = traceback)

runComputation("CommDataOrig.csv","CommDataOrig.csv", "results/r_orig.json", "results/r_orig_plot.png")
print("Done with original ---------------------------------------")
runComputation("SDV/datasets/syn_dataset.csv", "SDV/datasets/syn_dataset.csv", "results/r_sdv.json", "results/r_sdv_plot.png")
print("Done with SDV ---------------------------------------")
runComputation("ARX/datasets/syn_dataset.csv", "ARX/datasets/syn_dataset.csv", "results/r_arx.json", "results/r_arx_plot.png")
print("Done with ARX ---------------------------------------")
runComputation("synDiffix/datasets/sdx_toSchool_target_VO2max.csv", "synDiffix/datasets/sdx_toHome_target_VO2max.csv", "results/r_sdx.json", "results/r_sdx_plot.png")
print("Done with SynDiffix ---------------------------------------")