Cispa lecture statistics

examples/cispa-lecture-attendance.py

@@ -0,0 +1,92 @@
+#!/usr/bin/env python3
+import json
+import os
+from collections import defaultdict
+
+import cakecms
+
+HOST = 'https://cms.cispa.saarland'
+TOKEN = os.environ['CMS_TOKEN']
+TEMPFILE = '/tmp/cms.json'
+
+cms = cakecms.CakeCMS(HOST, TOKEN)
+# [optional] log queries to console
+#cms.debug = True
+# get a course list (dictionary)
+
+if os.path.exists(TEMPFILE):
+    # read data from cached file
+    with open(TEMPFILE, 'r') as f:
+        data = json.loads(f.read())
+    cysec1_courses = data['cysec1_courses']
+    security_courses = data['security_courses']
+    advanced_lectures = data['advanced_lectures']
+    cysec1_students = set(data['cysec1_students'])
+    security_students = set(data['security_students'])
+    advanced_lecture_students = {k: set(v) for k, v in data['advanced_lecture_students'].items()}
+else:
+    # read data from CMS
+    courses = cms.courses_index()
+    cysec1_courses = [course for course in courses if course['Course']['name'].strip() in ('Grundlagen der Cybersicherheit 1', 'Foundations of Cybersecurity 1')]
+    security_courses = [course for course in courses if course['Course']['name'].strip() == 'Security']
+    advanced_lectures = [course for course in courses if course['Course']['course_type'] == 'Advanced Lecture']
+
+    print('Cysec1:', [c['Course']['shortname'] for c in cysec1_courses])
+    print('Security:', [c['Course']['shortname'] for c in security_courses])
+    print('Advanced Lectures:', [c['Course']['shortname'] for c in advanced_lectures])
+
+    cysec1_students = set()  # set of matriculations
+    for course in cysec1_courses:
+        studis = cms.students_index(course=course['Course']['shortname'])['students']
+        cysec1_students |= set(student['Student']['matriculation'] for student in studis)
+
+    security_students = set()  # excluding cysec students
+    for course in security_courses:
+        studis = cms.students_index(course=course['Course']['shortname'])['students']
+        security_students |= set(student['Student']['matriculation'] for student in studis if student['Student']['subject'] not in ('Cybersicherheit', 'Entrepreneurial Cybersecurity'))
+
+    advanced_lecture_students = defaultdict(set)  # lecture name -> set of matriculations
+    for course in advanced_lectures:
+        studis = cms.students_index(course=course['Course']['shortname'])['students']
+        advanced_lecture_students[course['Course']['name']] |= set(student['Student']['matriculation'] for student in studis)
+    with open(TEMPFILE, 'w') as f:
+        f.write(json.dumps({
+            'cysec1_courses': cysec1_courses,
+            'security_courses': security_courses,
+            'advanced_lectures': advanced_lectures,
+            'cysec1_students': list(cysec1_students),
+            'security_students': list(security_students),
+            'advanced_lecture_students': {k: list(v) for k, v in advanced_lecture_students.items()},
+        }))
+
+# do your data analysis here!
+print('Cybersecurity 1 (all iterations)        :', len(cysec1_students), 'students')
+print('Security (all iterations)               :', len(security_students), 'students that do not study cybersecurity')
+print('Cybersecurity 1 /\\ Security             :', len(security_students & cysec1_students), 'students')
+print('-'*60)
+for lecture, matriculations in advanced_lecture_students.items():
+    print(f'{lecture:40s}: {len(matriculations)} students')
+
+print('-'*60)
+cysec_students_without_advanced_lecture = set(cysec1_students)
+security_students_without_advanced_lecture = set(security_students)
+for lec in advanced_lecture_students.values():
+    cysec_students_without_advanced_lecture -= lec
+    security_students_without_advanced_lecture -= lec
+print(len(cysec_students_without_advanced_lecture), 'cysec1 students did not attend any advanced lecture')
+print(len(security_students_without_advanced_lecture), 'security students did not attend any advanced lecture')
+
+#count advanced lectures per student
+for name, basic_set in [('cysec1', cysec1_students), ('security', security_students)]:
+    print('-'*60)
+    matriculation_to_advanced_count = {m: 0 for m in basic_set}
+    for students in advanced_lecture_students.values():
+        for m in students:
+            if m in matriculation_to_advanced_count:
+                matriculation_to_advanced_count[m] += 1
+    # statistics
+    cysec_advanced_count = defaultdict(lambda: 0)
+    for m, count in matriculation_to_advanced_count.items():
+        cysec_advanced_count[count] += 1
+    for count, num_students in sorted(cysec_advanced_count.items(), key=lambda a: a[0]):
+        print(f'{num_students:3d} {name} students took >> {count:2d} << advanced lectures')