spatial-group-detection/clustering.py at master · tillwf/spatial-group-detection · GitHub

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
# encoding: UTF-8

from metrics import haversine_acc
import numpy as np
import pandas as pd
from sklearn.cluster import DBSCAN
from sklearn.metrics.pairwise import pairwise_distances


def get_cluster_algorithm(algorithm):
    if algorithm == "db":
        """
        optimal eps values :
                "haversine": 0.05
                "haversine_acc": 0.05
        """
        return DBSCAN(
                    eps=0.05,
                    min_samples=2,
                    algorithm='brute',
                    metric=lambda X, Y: haversine_acc(X, Y)
                )


def normalize(clustering, time):
    from_minus1 = events_from_minus1(clustering, time)
    to_minus1 = events_to_minus1(clustering, time)

    return from_minus1 + to_minus1


def emit_event(clustering, time, previous_users, current_users, previous_cluster_id, current_cluster_id):
    event = {"date": time}

    if len(current_users) < 2:
        if len(previous_users):
            # Cluster deletion
            event.update({
                "type": "deletion",
                "cluster_id": previous_cluster_id,
                "people_out": previous_users
            })
        else:
            return
    else:
        mean_lat = clustering.loc[list(current_users)].latitude
        mean_long = clustering.loc[list(current_users)].longitude
        event.update({
            "centroid": (np.mean(mean_lat), np.mean(mean_long)),
            "population": current_users,
            "cluster_id": current_cluster_id,
            "density": mean_distance_intra_centroid(clustering, current_users)
        })

        if not len(previous_users):
            # Creation
            event.update({"type": "creation"})
        elif len(previous_users) > len(current_users):
            event.update({
                "type": "decrease",
                "previous_cluster_id": previous_cluster_id,
                "people_out": previous_users.difference(current_users)
            })

        elif len(previous_users) < len(current_users):
            event.update({
                "type": "increase",
                "previous_cluster_id": previous_cluster_id,
                "people_in": current_users.difference(previous_users)
            })
        else:
            print "Error !"

    print event
    return event


def mean_distance_intra_centroid(clustering, users):
    X = np.array(clustering.loc[users, ['latitude', 'longitude', 'accuracy']])
    try:
        distances = pairwise_distances(X, metric=lambda X, Y: haversine_acc(X, Y))
    except:
        distances = -1
    return np.mean(distances)


def events_from_minus1(clustering, time):

    prev_clus = clustering['clustering%s' % (time - 1)]
    current_clus = clustering['clustering%s' % time]

    # Handle previous non clustered points
    minus1 = pd.concat([prev_clus[prev_clus == -1], current_clus], axis=1, join='inner')
    new_groups = minus1[minus1['clustering%s' % time] != minus1['clustering%s' % (time - 1)]]

    events = []

    for cluster_id in set(new_groups['clustering%s' % time]):

        current_lines = clustering[clustering['clustering%s' % time] == cluster_id]
        current_users = set(current_lines.index.tolist())

        previous_lines = current_lines[current_lines['clustering%s' % (time - 1)] != -1].dropna()
        previous_users = set(previous_lines.index)

        previous_cluster_id = None
        if len(set(previous_lines.index)):
            # New users
            previous_cluster_id = list(set(current_lines['clustering%s' % (time - 1)].values) - set([-1]))[0]

        event = emit_event(clustering, time, previous_users, current_users, previous_cluster_id, cluster_id)
        events.append(event)

    return events


def events_to_minus1(clustering, time):

    prev_clus = clustering['clustering%s' % (time - 1)].dropna()
    current_clus = clustering['clustering%s' % time]

    # Handle previous non clustered points
    minus1 = pd.concat([prev_clus[current_clus == -1], current_clus], axis=1, join='inner')
    new_groups = minus1[minus1['clustering%s' % time] != minus1['clustering%s' % (time - 1)]]

    events = []

    for cluster_id in set(new_groups['clustering%s' % (time - 1)]):

        previous_lines = clustering[clustering['clustering%s' % (time - 1)] == cluster_id]
        previous_users = set(previous_lines.index.tolist())

        current_lines = new_groups[new_groups['clustering%s' % (time - 1)] == cluster_id]
        current_users = set(current_lines.index.tolist())

        kept_users = previous_users.difference(current_users)

        new_cluster_id = None

        if len(kept_users):
            # Cluster is alive
            users_in_cluster = current_clus.reset_index().groupby('clustering%s' % time).user_id.apply(set)
            try:
                new_cluster_id = users_in_cluster[users_in_cluster == kept_users].index.tolist()[0]
            except:
                print "There is a point that leaves with another point that arrives... problematic"
                new_cluster_id = -1

        event = emit_event(clustering, time, previous_users, kept_users, cluster_id, new_cluster_id)

        if event:
            events.append(event)

    return events