Class: Statistics

64
65
66
67
68
69
70
71
72
73

# File 'lib/statistics.rb', line 64

def contacts_sharing_with_count_sql
  <<SQL
    SELECT users.id AS id, count(contacts.id) AS count
      FROM users
        JOIN contacts on contacts.user_id = users.id
        JOIN aspect_memberships on aspect_memberships.contact_id = contacts.id
        #{self.where_clause_sql}
        GROUP BY users.id
SQL
end

104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119

# File 'lib/statistics.rb', line 104

def correlate(first_metric, second_metric)

  # [{"id" => 1 , "count" => 123}]

  x_array = []
  y_array = []

  self.result_hash(first_metric).keys.each do |k|
    if val = self.result_hash(second_metric)[k]
      x_array << self.result_hash(first_metric)[k]
      y_array << val
    end
  end

  correlation(x_array, y_array)
end

130
131
132
133
134
135
136
137
138
139
140
141
142
143

# File 'lib/statistics.rb', line 130

def correlation(x_array, y_array)
  sum = 0.0
  x_array.each_index do |i|
    sum = sum + x_array[i].to_f * y_array[i].to_f
  end
  x_y_mean = sum/x_array.length.to_f
  x_mean = mean(x_array)
  y_mean = mean(y_array)

  st_dev_x = standard_deviation(x_array)
  st_dev_y = standard_deviation(y_array)

  (x_y_mean - (x_mean*y_mean))/(st_dev_x * st_dev_y)
end

86
87
88
89
90
91
92
93
94

# File 'lib/statistics.rb', line 86

def fb_connected_distribution
  User.connection.select_all(fb_connected_distribution_sql).map { |row|
    Hash[
      row.map { |k,v|
        [k, v.to_i]
      }
    ]
  }
end

75
76
77
78
79
80
81
82
83
84

# File 'lib/statistics.rb', line 75

def fb_connected_distribution_sql
  <<SQL
    SELECT users.id AS id, users.sign_in_count AS count, count(services.id) AS connected
      FROM users
        LEFT OUTER JOIN services on services.user_id = users.id
          AND services.type = 'Services::Facebook'
        #{self.where_clause_sql}
        GROUP BY users.id, users.sign_in_count
SQL
end

121
122
123
124
125
126
127
128

# File 'lib/statistics.rb', line 121

def generate_correlations
  result = {}
  [:posts_count, :comments_count, :invites_sent_count, #:tags_followed_count,
   :mentions_count, :contacts_sharing_with_count].each do |metric|
    result[metric] = self.correlate(metric,:sign_in_count)
   end
  result
end

33
34
35
36
37
38
39
40
41

# File 'lib/statistics.rb', line 33

def invites_sent_count_sql
  <<SQL
    SELECT users.id AS id, count(invitations.id) AS count
      FROM users
        LEFT OUTER JOIN invitations ON users.id = invitations.sender_id
        #{self.where_clause_sql}
        GROUP BY users.id
SQL
end

145
146
147
148
149
150

# File 'lib/statistics.rb', line 145

def mean(array)
  sum = array.inject(0.0) do |sum, val|
    sum += val.to_f
  end
  sum / array.length
end

53
54
55
56
57
58
59
60
61
62

# File 'lib/statistics.rb', line 53

def mentions_count_sql
  <<SQL
    SELECT users.id AS id, count(mentions.id) AS count
      FROM users
        JOIN people on users.id = people.owner_id
        LEFT OUTER JOIN mentions on people.id = mentions.person_id
        #{self.where_clause_sql}
        GROUP BY users.id
SQL
end

11
12
13
14
15
16
17
18
19
20

# File 'lib/statistics.rb', line 11

def posts_count_sql
  <<SQL
    SELECT users.id AS id, count(posts.id) AS count
      FROM users
        JOIN people ON people.owner_id = users.id
        LEFT OUTER JOIN posts ON people.id = posts.author_id
        #{self.where_clause_sql}
        GROUP BY users.id
SQL
end

164
165
166

# File 'lib/statistics.rb', line 164

def retention(n)
  users_by_week(n).count.to_f/week_created(n).count
end

96
97
98
99
100
101
102

# File 'lib/statistics.rb', line 96

def sign_in_count_sql
  <<SQL
    SELECT users.id AS id, users.sign_in_count AS count
      FROM users
      #{self.where_clause_sql}
SQL
end

152
153
154
155
156
157
158
159
160
161

# File 'lib/statistics.rb', line 152

def standard_deviation(array)
  variance = lambda do
    m = mean(array)
    sum = 0.0
    array.each{ |v| sum += (v.to_f-m)**2 }
    sum/array.length.to_f
  end.call

  Math.sqrt(variance)
end

43
44
45
46
47
48
49
50
51

# File 'lib/statistics.rb', line 43

def tags_followed_count_sql
  <<SQL
    SELECT users.id AS id, count(tag_followings.id) AS count
      FROM users
        LEFT OUTER JOIN tag_followings on users.id = tag_followings.user_id
        #{self.where_clause_sql}
        GROUP BY users.id
SQL
end

168
169
170
171

# File 'lib/statistics.rb', line 168

def top_active_users(n)
  ten_percent_lim = (users_by_week(n).count.to_f * 0.3).ceil
  users_by_week(n).joins(:person => :profile).where('users.sign_in_count > 4').order("users.sign_in_count DESC").limit(ten_percent_lim).select('users.email, users.username, profiles.first_name, users.sign_in_count')
end

173
174
175

# File 'lib/statistics.rb', line 173

def users_by_week(n)
  week_created(n).where("current_sign_in_at > ?", Time.now - 1.week)
end