Easily form random groups of students for homework assignments. Aimed at Cornell faculty/staff using CMSX for course management.

hw_groups.py

#!/usr/bin/env python3
import pandas as pd
import numpy as np
GROUPSIZE = 4
INFILE = 'CS_5435_student_table.csv'
OUTFILE = 'CS_5435_hw_groups.csv'

''' 
    Load the student data (download it from
    https://cmsx.cs.cornell.edu/web/auth/?action=exporttable&courseid=<YOURCOURSEID>)
    Or click on 'Students' -> 'enrolled student table' on CMSX.
'''
df =  pd.read_csv(INFILE)
num_students = df.shape[0]

# randomize ordering of students 
df = df.sample(frac=1)

# group the students by GROUPSIZE
hw_groups = df.groupby(np.arange(num_students)//GROUPSIZE, group_keys=False)

# transpose the dataframe by netIDs to form groups
df = hw_groups.apply(lambda hw_group: hw_group['NetID'].reset_index(drop=True).to_frame().T)

# rename the transposed columns and reset the index
df.index.name = 'Group'
df.columns = ["Member"+str(x) for x in range(GROUPSIZE)]
df = df.reset_index(level=0, drop=True)

# check for groups without teammates (class not divisable by GROUPSIZE)
missing = df.apply(lambda group_members: group_members.isnull().sum(), axis=1)
if missing.sum() > 0:
    missed = missing[missing > 0]
    print("Group number {} (currently only {}) is missing {} member(s).\n"\
            "Group sizes of {} aren't evenly divisable by {}."\
            .format(missed.index[0], df.loc[missed.index[0],:].dropna().tolist(),\
            missed.values[0], GROUPSIZE, num_students))
    df = df.fillna('no member')

# save the group assignments to csv.
df.to_csv(OUTFILE, index=True)

print("The groups are as follows (saved to '{}'):".format(OUTFILE))
print('-'*80+"\n",df,"\n"+'-'*80)
print("Run this again to resample the groups.")

The input (roster) has the following columns for all Cornell courses:

['Last Name', 'First Name', 'NetID', ['Assignment names'],'CP', 'Final', 'Total Score', 'Final Grade']

As long as your roster is stored in CSV and has some column for NetID/email id or names of the students, just replace "NetID" on the gist to whatever your primary column is named.