In skranz/SeminarMatching: Shiny App for Seminar Assignment

knitr::opts_chunk$set(error = TRUE, echo=FALSE, cache=FALSE)
library(ggplot2)
library(SeminarMatching)
library(dplyr)
library(tidyr)
library(reshape2)

setwd("D:/libraries/SeminarMatching/semapps/shared/reports")
semester = "SS17"
db.dir = paste0("../db")
semid = 17
semdb = dbConnect(dbname=paste0(db.dir,"/semDB.sqlite"), drv = SQLite())

studpref = dbGet(semdb,"studpref",nlist(semester, semid))
if (NROW(studpref)==0) {
  stop("Cannot show seminar report, since no student has yet put the seminar in its preference.",call. = FALSE)
}

seminars = dbGet(semdb,"seminars",nlist(semester, semid))
students = dbGet(semdb,"students",nlist(semester))

df = studpref
cols = c("semid",setdiff(colnames(seminars),colnames(df)))
df = left_join(df, seminars[,cols,drop=FALSE], by="semid")
cols = c("userid",setdiff(colnames(students),colnames(df)))
df = left_join(df, students[,cols,drop=FALSE], by="userid")
seminar = as.list(seminars[1,])


num.prefs = NROW(df)

`r seminar$semname` `r semester`

A total of r num.prefs students have put the seminar in their preference list. Your seminar has r seminar$slots slots.

The following table shows how many students ranked the seminar at position pos.

df$pos = ifelse(df$pos <=3, df$pos, "4 or larger")
sdf = df %>%
  group_by(pos) %>% 
  summarise(n=n()) %>%
  arrange(pos)

sdf

The following table compares students that put the seminar in their preference list by their main subject and randked position

sdf = df %>%
  group_by(studSubject,pos) %>% 
  summarise(n=n()) %>%
  spread(key = pos, value=n) %>%
  replace.na(0, cols=-1) %>%
  as.data.frame()

sdf$total = rowSums(sdf[,-1,drop=FALSE],na.rm = TRUE)
sdf = arrange(sdf, -total) %>% select(studSubject, total, everything())

sdf

The following table compares WiWi students that put the seminar in their preference list by their specialization (students with 2 specializations count 1/2 for each specialization):

d = select(df,userid, studBAMA, studSpecBA, studSpecMA, studSpecMA2, pos)
d = tidyr::gather(d,key = specid,value = spec,studSpecBA, studSpecMA, studSpecMA2) %>%
  filter(spec!="-")


d = filter(d, 
  (studBAMA=="Bachelor" & specid=="studSpecBA") | 
  (studBAMA=="Master" & specid=="studSpecMA") |
  (studBAMA=="Master" & specid=="studSpecMA2")
)

d = d %>% 
  group_by(userid) %>%
  mutate(weight=1/n())


sdf = d %>%
  group_by(spec,pos) %>% 
  summarise(n=sum(weight)) %>%
  spread(key = pos, value=n) %>%
  replace.na(0, cols=-1) %>%
  as.data.frame()

sdf$total = rowSums(sdf[,-1, drop=FALSE],na.rm = TRUE)
sdf = arrange(sdf, -total) %>% select(spec, total, everything())
sdf