-
Notifications
You must be signed in to change notification settings - Fork 5
/
college_flow.py
319 lines (262 loc) · 9.46 KB
/
college_flow.py
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
"""
Construct a major flow diagram.
"""
import pandas as pd
import numpy as np
import igraph
COLLEGES = [
['#1f77b4', 'I', 'Incoming Freshman'],
['#2ca02c', 'G', 'Graduate'],
['#7f7f7f', 'D', 'Drop Out'],
['#8c564b', 'UN', 'University (Provost)'],
['#9edae5', 'LA', 'Humanities & Social Sciences'],
['#17becf', 'SC', 'College of Science'],
['#98df8a', 'BU', 'School of Management'],
['#9467bd', 'VS', 'Volgenau School of Engr.'],
['#aec7e8', 'AR', 'Visual & Performing Arts'],
['#c5b0d5', 'HH', 'Health and Human Services'],
['#c49c94', 'E1', 'Education & Human Development'],
['#d62728', 'CA', 'Conflict Analysis & Resolution'],
['#e377c2', 'PP', 'School of Policy, Government and International Affairs'],
['#f7b6d2', 'LW', 'School of Law']
]
COLORS = [
'#1f77b4',
'#2ca02c',
'#7f7f7f',
'#8c564b',
'#9edae5',
'#17becf',
'#98df8a',
'#9467bd',
'#aec7e8',
'#c5b0d5',
'#c49c94',
'#d62728',
'#e377c2',
'#f7b6d2',
'#ff7f0e',
'#ff9896',
'#ffbb78'
]
def read_student_df(fname='nsf_student.csv'):
student_df = pd.read_csv(fname)
# Build a MultiIndex
tuples = list(zip(student_df.TERMBNR.values, student_df.id.values))
idx = pd.MultiIndex.from_tuples(tuples)
# Then reindex the student data frame using that index
sframe = pd.DataFrame(student_df.values, columns=student_df.columns,
index=idx)
# Get all unique student ids and sort them, to build an idmap
student_ids = student_df['id'].unique()
student_ids.sort()
idmap = {sid: idx for idx, sid in np.ndenumerate(student_ids)}
# Now add the mapping to the student data frame
mapping = pd.DataFrame.from_dict(idmap, orient='index')
mapping.columns = ('idx',)
students = pd.merge(sframe, mapping, left_on=('id',), right_index=True)
# del students['id']
del students['TERMBNR']
# remove duplicate indices, keeping first occurence
students = students.groupby(level=(0,1)).first()
return students
def label_students(student_df):
# Get all unique student ids and sort them, to build an idmap
student_ids = student_df['id'].unique()
student_ids.sort()
# Build a label matrix for the 16 terms + 1
terms = student_df['TERMBNR'].unique()
terms = [200910] + list(terms)
terms.sort()
num_labels = len(terms)
slabels = np.ndarray((len(student_ids), num_labels), dtype=np.object_)
slabels.fill('N') # not attending
for term_num, term in enumerate(terms):
for sid in student_ids:
try:
record = student_df.ix[term].ix[sid]
coll = record['PCOLL']
idx = record['idx']
slabels[idx][term_num] = coll
except KeyError: # no courses in term
pass
# Put labels in data frame and annotate with terms as columns and student
# ids as index
labeldf = pd.DataFrame(slabels)
labeldf.columns = terms
labeldf.index = student_ids
labeldf.to_csv('student-college-labels.csv')
return labeldf
def next_term(termbnr):
"""Return the next term as an int."""
term = str(termbnr)
year = int(term[:4])
season = term[4:]
if season == '70':
tstring = "%d%d" % (year+1, 10)
elif season == '40':
tstring = "%d%d" % (year, 70)
else:
tstring = "%d%d" % (year, 40)
return int(tstring)
def prev_term(termbnr):
"""Return the previous term as an int."""
term = str(termbnr)
year = int(term[:4])
season = term[4:]
if season == '10':
tstring = "%d%d" % (year-1, 70)
elif season == '40':
tstring = "%d%d" % (year, 10)
else:
tstring = "%d%d" % (year, 40)
return int(tstring)
def add_admissions_labels(labeldf, fname='nsf_admissions.csv'):
admiss = pd.read_csv(fname)
aterms = admiss.drop_duplicates('id')[['id', 'cohort', 'Application_College']]
aterms['id'] = aterms.id.values.astype(np.int)
for sid, term, coll in aterms.values:
try:
if labeldf.loc[sid][term] != coll:
labeldf.loc[sid][term] = coll
labeldf.loc[sid][prev_term(term)] = 'I'
except ValueError: # some term was unknown in admissions cohort
pass
def add_graduation_labels(labeldf, fname='nsf_degrees.csv'):
deg = pd.read_csv(fname)
gterms = deg.drop_duplicates('id')[['id', 'GRADTERM', 'degcoll']]
gterms['id'] = gterms['id'].values.astype(np.int)
# The label G indicates graduated
for sid, term, coll in gterms.values:
if labeldf.loc[sid][term] == 'N':
labeldf.loc[sid][term] = 'G'
else:
labeldf.loc[sid][next_term(term)] = 'G'
def process_record(sid, record, labeldf):
"""Scan through and annotate gaps after enrollment as either
NS (no summer) or B (break).
"""
enrolled = False
num_breaks = 0
for term, label in record.iteritems():
if label == 'G':
break
if num_breaks == 2:
labeldf.loc[sid][term] = 'D'
break
if not enrolled:
if label == 'I':
enrolled = True
else:
if label == 'N':
if str(term).endswith('40'):
labeldf.loc[sid][term] = 'NS'
else:
labeldf.loc[sid][term] = 'B'
num_breaks += 1
else: # label is one of the colleges
num_breaks = 0
def process_student(label_vector):
"""Read through `slabels` and return edges to add to the graph that
represent the changing status of the student. `slabels` is a vector of
labels, one for each term. The labels/nodes are 'I', 'G', or one of the 9
colleges. Note that this does not include 'NS', 'N', or 'B' as labels of
interest. We might at some point be interested in how often students from
each college take breaks.
"""
try:
start_index = label_vector.tolist().index('I')
except ValueError:
raise StopIteration()
cur_label = label_vector.ix[start_index]
idx = start_index
for label in label_vector[start_index:-1]:
if label == 'G' or label == 'D':
break
if label == 'N':
continue
if label == 'NS' or label == 'B':
label = cur_label
next_label = label_vector.ix[idx+1]
if next_label != 'NS' and next_label != 'B' and label != next_label:
yield (label, next_label)
cur_label = next_label
idx += 1
def yield_edges(student_label_df):
"""Iterate over each row of the student labels data frame, yielding an edge
for each college transition.
"""
for sid, slabel_vector in student_label_df.iterrows():
try:
edges = process_student(slabel_vector)
for edge in edges:
yield edge
except StopIteration:
pass
def build_college_flow_graph(slabels):
nodes = ('I','G','D', 'UN', 'LA', 'SC', 'BU', 'VS', 'AR', 'HH', 'E1', 'CA',
'PP', 'EI', 'IC', 'LW')
edges = yield_edges(slabels)
g = igraph.Graph(directed=True)
g.add_vertices(nodes)
g.add_edges(edges)
return g
def get_outweight(node):
succs = [v.index for v in node.successors()]
edge_pairs = zip([node.index]*len(succs), succs)
edge_ids = [g.get_eid(*edge) for edge in edge_pairs]
weights = [g.es[eid]['weight'] for eid in edge_ids]
return sum(weights)
def get_inweight(node):
preds = [v.index for v in node.predecessors()]
edge_pairs = zip(preds, [node.index]*len(preds))
edge_ids = [g.get_eid(*edge) for edge in edge_pairs]
weights = [g.es[eid]['weight'] for eid in edge_ids]
return sum(weights)
def prep_for_display(graph):
graph.simplify(combine_edges=sum)
graph.vs['inflow'] = map(get_inflow, graph.vs)
graph.vs['outflow'] = map(get_outweight, graph.vs)
return graph
def flow_matrix(graph):
# combine multiple edges, so that the resulting edges
# are weighted with the number of edges combined
if not graph.is_weighted():
graph.es['weight'] = 1
g = graph.simplify(combine_edges=sum)
else:
g = graph
# build the flow matrix, where cell (i, j) contains the
# number of edges that moved from node i to node j (edge weight)
num_nodes = len(graph.vs)
matrix = np.ndarray((num_nodes, num_nodes))
for v in graph.vs:
for s in v.successors():
eid = g.get_eid(v.index, s.index)
matrix[v.index][s.index] = g.es[eid]['weight']
# return the normalized matrix
return matrix / matrix.sum()
def write_college_csv(fname='colleges.csv'):
# TODO: add attributes to graph and write from graph
with open(fname, 'w') as f:
writer = csv.writer(f)
writer.writerow(('color', 'abbrev', 'name'))
rows = [','.join(row) for row in COLLEGES]
writer.writerows(rows)
def main():
student_df = read_student_df()
labeldf = label_students(student_df)
# Next we need data from admissions and graduation
add_admissions_labels(labeldf)
add_graduation_labels(labeldf)
# Next, scan through and annotate gaps after enrollment as either
# NS (no summer) or B (break)
for sid, record in labeldf.iterrows():
process_record(sid, record, labeldf)
# save the labels
labeldf.to_csv('student-college-labels.csv')
# Get graph edges
graph = build_college_flow_graph(labeldf)
graph.write_picklez('college-flow-graph.picklez')
if __name__ == "__main__":
sys.exit(main())