1 files changed, 71 insertions, 0 deletions
diff --git a/analyze_grades.r b/analyze_grades.r
new file mode 100755
index 0000000..6b64099
--- /dev/null
+++ b/analyze_grades.r
@@ -0,0 +1,71 @@
+#!/usr/bin/env Rscript
+# Copyright (C) 2018 Ryan Kavanagh <rkavanagh@cs.cmu.edu>
+#
+# Processes a CSV file into a format that Canvas will accept.
+# Also spits out statistics on the grades and pretty plots.
+
+library(matrixStats)
+library(tidyr)
+library(ggplot2)
+
+args = commandArgs(trailingOnly=TRUE)
+
+if(length(args)==0) {
+    stop("supply csv filename as cli arg")
+}
+
+HW <- args[1]
+
+metadata <- c("Student", "ID","Section", "SIS Login ID")
+
+hwdata <- read.csv(HW, check.names=FALSE)
+grades <- hwdata[, setdiff(names(hwdata), metadata)]
+totals <- rowSums(grades)
+hwdata["total"] <- totals
+grades <- hwdata[, setdiff(names(hwdata), metadata)]
+
+gradesMatrix = as.matrix(grades)
+
+stats <- data.frame( "mean" = colMeans(gradesMatrix, na.rm=TRUE)
+                  , "sd" = colSds(gradesMatrix, na.rm=TRUE)
+                  , "median" = colMedians(gradesMatrix, na.rm=TRUE)
+                  , "max" = colMaxs(gradesMatrix, na.rm=TRUE)
+                  , "min" = colMins(gradesMatrix, na.rm=TRUE))
+
+print("Statistics")
+print(stats)
+
+print("Raw anonymized grades")
+print(gradesMatrix[sample(nrow(gradesMatrix)),])
+
+myrange <- function(x) max(x) - min(x)
+
+                                        # Binwidth
+scottbw <- function(x) 3.49 * sd(x) * length(x)^(1/3)
+fdrbw <- function(x) 2 * IQR(x) / length(x)^(1/3)
+
+mybw <- function(x) {
+    if (myrange(x) <= 5) {
+        1
+    } else if (myrange(x) <= 10) {
+        2
+    } else {
+        myrange(x) / 6
+    }
+}
+
+pdf(paste(HW, "_plot.pdf", sep=""))
+ggplot(gather(grades), aes(value)) +
+    geom_histogram(binwidth = function(x) mybw(x)) +
+    facet_wrap(~key, scales = 'free_x') +
+    xlim(c(-1,NA))
+dev.off()
+
+if(length(args) == 2) {
+    output <- hwdata[metadata]
+    output["SIS Login ID"] <- NULL
+    output[args[2]] <- totals
+    write.csv(output, file = paste(HW, "_totals.csv", sep=""), na="", row.names=FALSE)
+}
+
+warnings()