#! /usr/bin/env python3
#
def homes ( ):

#*****************************************************************************80
#
## homes() makes a linear model of home selling prices.
#
#  Licensing:
#
#    This code is distributed under the GNU LGPL license. 
#
#  Modified:
#
#    21 August 2019
#
#  Author:
#
#    John Burkardt
#
  import matplotlib.pyplot as plt
  import numpy as np

  data = np.loadtxt ( 'homes_data.txt' )
#
#  Extract data items.
#
  ask = data[:,0]
  sel = data[:,1]
  liv = data[:,2]
  rms = data[:,3]
  bed = data[:,4]
  bat = data[:,5]
  age = data[:,6]
  lot = data[:,7]
  tax = data[:,8]
#
#  Count data items.
#
  n = len ( ask )
  print ( '  Number of listings is ', n )
#
#  Create matrix A = [1,liv,rms,bed,bat,age,lot,tax].
#
  A = np.zeros ( [ n, 8 ] )
  A[:,0] = np.ones ( n )
  A[:,1] = liv
  A[:,2] = rms
  A[:,3] = bed
  A[:,4] = bat
  A[:,5] = age
  A[:,6] = lot
  A[:,7] = tax

  cask, _, _, _ = np.linalg.lstsq ( A, ask, rcond = None )
#
#  Use A and c to evaluate the linear model for the bill.
#
  ask_predict = np.dot ( A, cask )
#
#  Print the formula
#
  print ( 'predicted asking price = ' )
  print ( '%g' % ( cask[0] ) )
  print ( '+ %g * living space' % ( cask[1] ) )
  print ( '+ %g * number of rooms' % ( cask[2] ) )
  print ( '+ %g * number of bedrooms' % ( cask[3] ) )
  print ( '+ %g * number of bathrooms' % ( cask[4] ) )
  print ( '+ %g * age in years' % ( cask[5] ) )
  print ( '+ %g * lot size in acres' % ( cask[6] ) )
  print ( '+ %g * annual taxes in dollars.' % ( cask[7] ) )
#
#  Plot the actual and predicted asking prices.
#
  askmin = min ( ask )
  askmax = max ( ask )
  mm = np.array ( [ askmin, askmax ] )
  plt.plot ( mm, mm, 'r-', linewidth = 3 )
  plt.plot ( ask, ask_predict, 'bo', markersize = 10 )
  plt.grid ( True )
  plt.xlabel ( 'Actual asking price' )
  plt.ylabel ( 'Predicted asking price' )
  plt.title ( 'Linear model of asking price for a house.' )
  filename = 'homes.png'
  plt.savefig ( filename )
  print ( '  Graphics saved as "', filename, '"' )
  plt.show ( )
#
#  Compute a selling price for a new listing.
#
  data2 = np.loadtxt ( 'homes_test.txt' )
#
#  Extract new data items.
#
  ask2 = data2[:,0]
  sel2 = data2[:,1]
  liv2 = data2[:,2]
  rms2 = data2[:,3]
  bed2 = data2[:,4]
  bat2 = data2[:,5]
  age2 = data2[:,6]
  lot2 = data2[:,7]
  tax2 = data2[:,8]
#
#  If only one new listing, force it to "look like" a vector.
#
  ask2 = np.atleast_1d ( ask2 )

  n2 = len ( ask2 )
  print ( '  Number of new listings is ', n2 )

  A2 = np.zeros ( [ n2, 8 ] )
  A2[:,0] = np.ones ( n2 )
  A2[:,1] = liv2
  A2[:,2] = rms2
  A2[:,3] = bed2
  A2[:,4] = bat2
  A2[:,5] = age2
  A2[:,6] = lot2
  A2[:,7] = tax2
#
#  Evaluate the linear model.
#
  ask2_predict = np.dot ( A2, cask )

  print ( '' )
  print ( '  Realtor   Model' )
  print ( '' )
  for i in range ( 0, n2 ):
    print ( '  %g  %g' % ( ask2[i], ask2_predict[i] ) )

  return

if ( __name__ == "__main__" ):
  homes ( )