Class: NyFrbAsDataframe::Hhdc

Inherits:
Object
  • Object
show all
Includes:
HTTParty
Defined in:
lib/ny_frb_as_dataframe/hhdc.rb

Instance Attribute Summary collapse

Instance Method Summary collapse

Constructor Details

#initialize(series = nil) ⇒ Hhdc

Returns a new instance of Hhdc.



17
18
19
 
# File 'lib/ny_frb_as_dataframe/hhdc.rb', line 17

def initialize(series = nil)
  @tag = series
end

Instance Attribute Details

#tag ⇒ Object (readonly)

Returns the value of attribute tag.



15
16
17
 
# File 'lib/ny_frb_as_dataframe/hhdc.rb', line 15

def tag
  @tag
end

Instance Method Details

#fetch(start: nil, fin: nil) ⇒ Object 



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
 
# File 'lib/ny_frb_as_dataframe/hhdc.rb', line 21

def fetch(start: nil, fin: nil)
  df = nil

  SimpleXlsxReader.configuration.auto_slurp = true

  Tempfile.open(['foo', '.xlsx'], '/tmp', binmode: true) do |f|
    qtr = Date.today.year * 4 + (Date.today.month.div(3)+1)
    dta = nil
    resp = nil

    loop do
      u = "https://www.newyorkfed.org/medialibrary/interactives/householdcredit/data/xls/hhd_c_report_#{qtr.div(4)}q#{qtr.modulo(4)+1}.xlsx"
      dta = self.class.get(u)
      resp = dta.parsed_response
      break if (dta.response.code.to_i == 200 && dta.parsed_response.encoding.to_s == "ASCII-8BIT")
      qtr -= 1
    end
    f.write(resp)
    f.rewind

    xlsx_file = SimpleXlsxReader.open(f.path)

    col_nm_hsh = 
      {
        'Total Debt Balance and Its Composition' => {'Mortgage' => 'Mort Debt Balance ($T)', 'HE Revolving' => 'HELOC Debt Balance ($T)',  'Auto Loan' => 'Car Loan Debt Balance ($T)', 'Credit Card' => 'CC Debt Balance ($T)', 'Student Loan' => 'Student Debt Balance ($T)',  'Other' => 'Other Debt Balance ($T)', 'Total' => 'Total Debt Balance ($T)'},
        'Number of Accounts by Loan Type' => {'Auto Loan' => 'Num Car Loan Accts (M)', 'Credit Card' => 'Num CC Accts (M)', 'Mortgage' => 'Num Mort Accts (M)',  'HE Revolving' => 'Num HELOC Accts (M)'},
        'Total Number of New and Closed Accounts and Consumer Credit Inquiries' => {'inquiry within 6 mo' => 'Num Credit Inquiries and New Accts', 'closed within 12 mo' => 'Num New Credit Accts Closed in 12 mo', 'open within 12 mo' => 'Num New Credit Accts Open after 12 mo'},
        'Mortgage Origination Volume by Riskscore' => {'<620' => 'New Mort w Riskscore <620 ($B)','620-659' => 'New Mort w Riskscore 620-659 ($B)','660-719' => 'New Mort w Riskscore 660-719 ($B)','720-759' => 'New Mort w Riskscore 720-759 ($B)','760+' => 'New Mort w Riskscore 760+ ($B)','TOTAL' => 'New Mort ($B)'},
        'Distribution of Riskscore of Mortgage Originations' => {'Median' => 'Median Riskscore for New Mort', '25th percentile' => 'Bottom 25%ile Riskscore for New Mort', '10th percentile' => 'Bottom 10%ile Riskscore for New Mort'},
        'Auto Loan Origination Volume by Riskscore' => {'<620' => 'New Car Loans w Riskscore <620 ($B)','620-659' => 'New Car Loans w Riskscore 620-659 ($B)','660-719' => 'New Car Loans w Riskscore 660-719 ($B)','720-759' => 'New Car Loans w Riskscore 720-759 ($B)','760+' => 'New Car Loans w Riskscore 760+ ($B)','TOTAL' => 'All New Car Loans ($B)'},
        'Distribution of Riskscore of Auto Loan Originations' => {'Median' => 'Median Riskscore for New Car Loan', '25th percentile' => 'Bottom 25%ile Riskscore for New Car Loan', '10th percentile' => 'Bottom 10%ile Riskscore for New Car Loan'},
        'Credit Limit and Balance for Credit Cards and HE Revolving' => {'Credit Card Balance' => 'CC Balance ($T)', 'Credit Card Available Credit' => 'CC Avail Credit ($T)',  'Credit Card Limit' => 'CC Limit ($T)', 'HE Revolving Balance' => 'HELOC Balance ($T)',  'HE Revolving Available Credit' => 'HELOC Avail Credit ($T)', 'HE Revolving Limit' => 'HELOC Limit ($T)'},
        'Total Balance by Delinquency Status' => {'Current' => '% Balances Current', '30 days late' => '% Balances 30d late',  '60 days late' => '% Balances 60d late',  '90 days late' => '% Balances 90d late',  '120+ days late' => '% Balances 120+d late',  'Severely Derogatory' => '% Balances Severely Impair'},
        'Percent of Balance 90+ Days Delinquent by Loan Type' => {'MORTGAGE' => '% Mort Bal 90d+ late',  'HELOC' => '% HELOC Bal 90d+ late', 'AUTO' => '% Car Loan Bal 90d+ late',  'CC' => '% CC Bal 90d+ late',  'STUDENT LOAN' => '% Student Loan Bal 90d+ late',  'OTHER' => '% Other Bal 90d+ late', 'ALL' => '% All Bal 90d+ late'},
        'New Delinquent* Balances by Loan Type' => {'AUTO' => '% Car Loan Bal 0d->30d+ Late',  'CC' => '% CC Bal 0d->30d+ Late',  'MORTGAGE' => '% Mort Bal 0d->30d+ Late',  'HELOC' => '% HELOC Bal 0d->30d+ Late', 'STUDENT LOAN' => '% Student Loan Bal 0d->30d+ Late',  'OTHER' => '% Other Bal 0d->30d+ Late', 'Total' => '% All Bal 0d->30d+ Late'},
        'New Seriously Delinquent* Balances by Loan Type' => {'AUTO' => '% Car Loan Bal ->90d+ Late',  'CC' => '% CC Bal ->90d+ Late',  'MORTGAGE' => '% Mort Bal ->90d+ Late',  'HELOC' => '% HELOC Bal ->90d+ Late', 'STUDENT LOAN' => '% Student Loan Bal ->90d+ Late',  'OTHER' => '% Other Bal ->90d+ Late', 'ALL' => '% All Bal ->90d+ Late'},
        'Quarterly Transition Rates for Current Mortgage Accounts' => {'To 30-60 days late' => 'Qtrly % Mort 0->30-60d late', 'To 90+ days Late' => 'Qtrly % Mort ->90d+ Late'},
        'Quarterly Transition Rates for 30-60 Day Late Mortgage Accounts' => {'To current'  => 'Qtrly % Mort % 30-60d->0d Late', 'To 90+ days Late' => 'Qtrly % Mort 30-60d->90d+ Late'},
        'Number of Consumers with New Foreclosures and Bankruptcies' => {'foreclosure' => 'Num Consumers w New Foreclosure',  'bankruptcy' => 'Num Consumers w New Bankruptcy'},
        'Third Party Collections' => {'proportion of consumers with collection' => '% Consumers w Collection', 'average collection amount per person with item' => 'Avg Collection Amt'},
        'Total Debt Balance by Age' => {'18-29' => 'Avg Total Debt Bal for 18-29 yo ($T)', '30-39' => 'Avg Total Debt Bal for 30-39 yo ($T)', '40-49' => 'Avg Total Debt Bal for 40-49 yo ($T)', '50-59' => 'Avg Total Debt Bal for 50-59 yo ($T)', '60-69' => 'Avg Total Debt Bal for 60-69 yo ($T)', '70+' => 'Avg Total Debt Bal for 70+ yo ($T)'},
        'Auto Loan Originations By Age' => {'18-29' => 'New Car Loans for 18-29 yo ($B)', '30-39' => 'New Car Loans for 30-39 yo ($B)', '40-49' => 'New Car Loans for 40-49 yo ($B)', '50-59' => 'New Car Loans for 50-59 yo ($B)', '60-69' => 'New Car Loans for 60-69 yo ($B)', '70+' => 'New Car Loans for 70+ yo ($B)'},
        'Total Mortgage Orginations by Age' => {'18-29' => 'New Mort for 18-29 yo ($B)', '30-39' => 'New Mort for 30-39 yo ($B)', '40-49' => 'New Mort for 40-49 yo ($B)', '50-59' => 'New Mort for 50-59 yo ($B)', '60-69' => 'New Mort for 60-69 yo ($B)', '70+' => 'New Mort for 70+ yo ($B)'},
        'Transition into Serious Delinquency (90+) by Age' => {'18-29' => 'Qtrly % ->90d+ late for 18-29 yo', '30-39' => 'Qtrly % ->90d+ late for 30-39 yo', '40-49' => 'Qtrly % ->90d+ late for 40-49 yo', '50-59' => 'Qtrly % ->90d+ late for 50-59 yo', '60-69' => 'Qtrly % ->90d+ late for 60-69 yo', '70+' => 'Qtrly % ->90d+ late for 70+ yo', 'all' => 'Qtrly % ->90d+ late'},
        'Transition into Serious Delinquency (90+) for Mortgages by Age' => {'18-29' => 'Qtrly % Mort ->90d+ late for 18-29 yo', '30-39' => 'Qtrly % Mort ->90d+ late for 30-39 yo', '40-49' => 'Qtrly % Mort ->90d+ late for 40-49 yo', '50-59' => 'Qtrly % Mort ->90d+ late for 50-59 yo', '60-69' => 'Qtrly % Mort ->90d+ late for 60-69 yo', '70+' => 'Qtrly % Mort ->90d+ late for 70+ yo', 'all' => 'Qtrly % Mort ->90d+ late'},
        'Transition into Serious Delinquency (90+) for Auto Loans by Age' => {'18-29' => 'Qtrly % Car Loans ->90d+ late for 18-29 yo', '30-39' => 'Qtrly % Car Loans ->90d+ late for 30-39 yo', '40-49' => 'Qtrly % Car Loans ->90d+ late for 40-49 yo', '50-59' => 'Qtrly % Car Loans ->90d+ late for 50-59 yo', '60-69' => 'Qtrly % Car Loans ->90d+ late for 60-69 yo', '70+' => 'Qtrly % Car Loans ->90d+ late for 70+ yo', 'all' => 'Qtrly % Car Loans ->90d+ late'},
        'Transition into Serious Delinquency (90+) for Credit Cards by Age' => {'18-29' => 'Qtrly % CC ->90d+ late for 18-29 yo', '30-39' => 'Qtrly % CC ->90d+ late for 30-39 yo', '40-49' => 'Qtrly % CC ->90d+ late for 40-49 yo', '50-59' => 'Qtrly % CC ->90d+ late for 50-59 yo', '60-69' => 'Qtrly % CC ->90d+ late for 60-69 yo', '70+' => 'Qtrly % CC ->90d+ late for 70+ yo', 'all' => 'Qtrly % CC ->90d+ late'},
        'Transition into Serious Delinquency (90+) for Student Loans by Age' => {'18-29' => 'Qtrly % Student Loan ->90d+ late for 18-29 yo', '30-39' => 'Qtrly % Student Loan ->90d+ late for 30-39 yo', '40-49' => 'Qtrly % Student Loan ->90d+ late for 40-49 yo', '50-59' => 'Qtrly % Student Loan ->90d+ late for 50-59 yo', 'all' => 'Qtrly % Student Loan ->90d+ late'},
        'New Foreclosures By Age' => {'18-29' => 'New Foreclosures for 18-29 yo (000)', '30-39' => 'New Foreclosures for 30-39 yo (000)', '40-49' => 'New Foreclosures for 40-49 yo (000)', '50-59' => 'New Foreclosures for 50-59 yo (000)', '60-69' => 'New Foreclosures for 60-69 yo (000)', '70+' => 'New Foreclosures for 70+ yo (000)'},
        'New Bankruptcy By Age' => {'18-29' => 'New Bankruptcies for 18-29 yo (000)', '30-39' => 'New Bankruptcies for 30-39 yo (000)', '40-49' => 'New Bankruptcies for 40-49 yo (000)', '50-59' => 'New Bankruptcies for 50-59 yo (000)', '60-69' => 'New Bankruptcies for 60-69 yo (000)', '70+' => 'New Bankruptcies for 70+ yo (000)'}
      }

    # Polars::Config.set_tbl_cols(-1)
    Polars::Config.set_tbl_rows(-1)

    done = false
    xlsx_file.sheets.each do |sht|  # Begin per-sheet analysis
      nm = sht.name

      begin 
        _ = sht.rows
        pg_nm_text = sht.rows
        pg_nm_text = pg_nm_text.first 
      rescue 
        begin
          pg_nm_text = sht.rows.first
        rescue 
          done = true
        end
      end

      break if done == true

      pg_nm_text = pg_nm_text.first
      next unless col_nm_hsh.keys.include?(pg_nm_text)

      header_found = false
      csv_txt = []

      sht.rows.each do |row|   # Begin per-line analysis
        r = row

        if header_found == false 
          if r.length > (r - col_nm_hsh[pg_nm_text].keys).length
            header_found = true 
            0.upto(r.length-1).to_a.each do |ndex|
              if ndex.zero?
                r[ndex] = 'Timestamps_'
              else
                col_nm_hsh[pg_nm_text].keys.each do |k|
                  r[ndex] = col_nm_hsh[pg_nm_text][k] if r[ndex] == k
                end
              end
            end
          end

          next if header_found == false
        end 

        break if r.compact.length.zero?

        # puts "#{__FILE__}:#{__LINE__} r = >#{r.inspect}<"
        if /(?<yr>\d{2}):Q(?<qtr>\d)/ =~ r[0].to_s
          yr = yr.to_i + 2000
          yr -= 100 if yr > Date.today.year
          dte = Date.new(yr, qtr.to_i * 3 - 2, 1)
          r[0] = dte
        elsif (r[0].to_date rescue false)
          r[0] = r[0].to_date.beginning_of_quarter
        end

        next if r.map{|rr| (rr.nil? ? '' : rr.to_s).length }.sum.zero?
        rr = r.map{|c| c.nil? ? '' : c}
        rr = rr.reverse.drop_while(&:nil?).reverse
        next if rr.compact.length.zero?
        csv_txt << rr.join(',')

      end    # End per-line analysis

      csv_dta = CSV.parse(csv_txt.join("\n"), headers: true, converters: [:numeric])
      csv_dta.delete(nil)

      keys = csv_dta.headers.compact.intersection(['Timestamps_'] + col_nm_hsh[pg_nm_text].values)
      vals = csv_dta.to_a[1..-1].map{|r| (r + [''] * keys.length)[0..keys.length-1] }.transpose

      tmp_df = {}; (0..(keys.length-1)).to_a.each{|i| tmp_df[keys[i]] = vals[i]}

      tmp_df = Polars::DataFrame.new(tmp_df)
      tmp_df = tmp_df.filter(Polars.col('Timestamps_') <= fin.try(:to_date)) unless fin.nil?
      tmp_df = tmp_df.filter(Polars.col('Timestamps_') >= start.try(:to_date)) unless start.nil?
      tmp_df.sort!('Timestamps_')

      if df.nil?
        df = tmp_df.dup 
      else 
        df = df.join(tmp_df, on: 'Timestamps_', how: 'full')

        tmp_timestamps = df['Timestamps_'].to_a 
        tmp_timestamps__ = df['Timestamps__right'].to_a
        tmp_timestamps.length.times.each {|ndex| tmp_timestamps[ndex] ||= tmp_timestamps__[ndex] }
        df.replace('Timestamps_', Polars::Series.new(tmp_timestamps))
        df = df.drop(['Timestamps__right'])
      end
      df = df.sort('Timestamps_')
 
    end    # End per-sheet analysis

    s = Polars::Series.new('Timestamps', df['Timestamps_'].to_a)
    df = df.insert_column(0,s).drop(['Timestamps_'])
  end

  df = df.drop(df.columns - ['Timestamps', @tag]) unless @tag.nil? || !df.columns.include?(@tag)
  df
end