Treemap of Aalto courses by school and department

gistfile1.r

####################################################
#
# Treemap of Aalto courses by school and department.
# The color palette indicates the average ECTS
# of the courses of the department.
#
# Data from Linked Open Aalto Data Service
# http://data.aalto.fi/
#
# Tuija Sonkkila 27.11.2012, 18.5.2013
#
# Note that ECON courses do not have
# any granularity 
#
# The official school acronyms of today are
# TaiK -> ARTS
# ECON -> BIZ
# CHEM
# ELEC
# ENG
# SCI
#
####################################################

library(SPARQL)

endpoint <- "http://data.aalto.fi/sparql"

q <- '
SELECT ?courseCode ?ectsValue ?deptCode ?schoolCode 
WHERE 
{
  GRAPH <http://data.aalto.fi/id/courses/noppa/> 
  {
   ?course <http://purl.org/vocab/aiiso/schema#code> ?courseCode ;
           <http://linkedscience.org/teach/ns#ects> ?ects .
   BIND (IF (regex(?ects, "-"),           
            REPLACE(?ects, "^.*-", ""), 
            ?ects) AS ?ectsValue)
  
   ?dept <http://purl.org/vocab/aiiso/schema#teaches> ?course ;
         <http://purl.org/vocab/aiiso/schema#code> ?deptCode ;
         <http://purl.org/vocab/aiiso/schema#part_of> ?school .
  
   ?school <http://purl.org/vocab/aiiso/schema#organization> ?dept ;
           <http://purl.org/vocab/aiiso/schema#code> ?schoolCode .
 }
} 
ORDER BY ?schoolCode ?deptCode ?courseCode'

res <- SPARQL(url=endpoint, query=q)$results

# Take a copy of the res dataframe
coursedata <- res

# Change the datatype of ects
coursedata$ectsValue <- as.numeric(coursedata$ectsValue)

# EDIT 18.5.2013 Rename ECON -> BIZ, TaiK -> ARTS, Eri -> OU (Open University)
coursedata$schoolCode <- sub('ECON', 'BIZ', coursedata$schoolCode)
coursedata$schoolCode <- sub('TaiK', 'ARTS', coursedata$schoolCode)
coursedata$schoolCode <- sub('ERI', 'OU', coursedata$schoolCode)

# Factorize dept names
#
# 29.11.2012 don't do this, results in a malplotted map
# see https://blogs.aalto.fi/suoritin/2012/11/28/course-data-up-on-a-treemap/
# coursedata$deptCode <- factor(coursedata$deptCode)

# Count number of courses per department
deptsCourses <- aggregate(courseCode ~ deptCode,
                          data = coursedata,
                          FUN = "length")

# Calculate average ects by department with aggregating by department and school
depts <- aggregate(x = coursedata$ectsValue, 
                   by = list(dept = coursedata$deptCode, 
                             school = coursedata$schoolCode), 
                   FUN = mean)

# Merge the two dataframes by the department code
coursedata.merged <- merge(x = deptsCourses, 
                           y = depts, 
                           by.x = "deptCode",
                           by.y = "dept")

# Clean up variable names
colnames(coursedata.merged) <- c("dept", "courses", "school", "ECTS average")

# Plot a treemap with the treemap library and with a certain Brewer palette
library(treemap)
library(RColorBrewer)

png("aalto.courses.treemap.png", width = 1024, height = 768, res = 72)
tmPlot(coursedata.merged,
       index = c("school", "dept"),
       vSize = "courses",
       vColor = "ECTS average",
       type = "value",
       palette = "YlOrBr",    
       range = c(0, 11),
       title = "Treemap of Aalto courses by School and department",
       subtitle = "Average ECTS by department")
dev.off()

# A treemap with the portfolio library. The color is hard-coded though
#
# library(portfolio)
#
# png("aalto.courses.portfolio.treemap.png", width = 1024, height = 768, res = 72)
# map.market(id = coursedata.merged$dept, 
#           area = coursedata.merged$courses, 
#           group = coursedata.merged$school, 
#           color = coursedata.merged$'ECTS average', 
#           lab = c(TRUE, TRUE),
#           main="Aalto courses by School")
# dev.off()