[continued from previous message]   
      
   Avg Temp^2/1000            -1.3258       0.0513     -25.8514  0.0000   
   Avg Temp^3/1000             0.0130       0.0007      19.3185  0.0000   
   Winter Days Away>9         -0.0088       0.0018      -4.9382  0.0000   
   Compressor Failure Days     0.0214       0.0028       7.7401  0.0000   
   New Windows 2020           -0.1103       0.0243      -4.5320  0.0000   
   Porch Enclosed 2014        -0.1262       0.0238      -5.3112  0.0000   
   New Water Heater 2017      -0.0652       0.0278      -2.3434  0.0202   
   New Heat Pump 2024         -0.2033       0.0313      -6.4965  0.0000   
   May                        -0.1674       0.0424      -3.9500  0.0001   
   Jun                        -0.1616       0.0646      -2.5028  0.0132   
   Jul                        -0.1265       0.0783      -1.6148  0.1081   
   Aug                        -0.1707       0.0720      -2.3726  0.0187   
   Sep                        -0.1653       0.0581      -2.8465  0.0049   
   Oct                        -0.1464       0.0366      -3.9997  0.0001   
   September Outliers         -0.2115       0.0608      -3.4781  0.0006   
      
      
   I was very interested in seeing if new heat pump is more efficient than   
   the 15 year old unit it replaced. It is, and the result is highly   
   significant with a t stat of 6.4965.   
      
   Plot highlighting new heat pump months:   
      
   https://drive.google.com/file/d/1oKlT9dM3pztd2AcI0wYG6ATIXRysLtF   
   /view?usp=sharing   
      
   Note that for any given temperature the green dots for the months with   
   the new heat pump are all on the low side of the other data points. Many   
   lie below any dots at that temperature. Thus the highly significant t Stat.   
      
   We made other post-2010 improvements and repairs too. Roof replaced,   
   exterior paint (twice), kitchen appliances replaced, master bath   
   remodel, one exterior door replaced, water-damaged chimney casing   
   repaired, ceiling fans added to all 4 bedrooms, all come to mind. No   
   reason to suspect anything other than the appliances would make a   
   difference. Appliances did, but not statistically significant. The   
   exterior door is a garage entry, not a heated area.   
      
   Finally, the November regression residual is -0.0122, only 0.3% of the   
   actual value. The model continues to track the actual use for another   
   month. Note that I moved the vertical axis from 0 to 2.5 to highlight   
   the differences between actual and predicted.   
      
   https://drive.google.com/file/d/1b1wKGpxpPbuKgRmgUT1xG9UGGy8HVvU   
   /view?usp=sharing   
      
      
   Source: Average Temperature: https://www.weather.gov/wrh/climate?wfo=ind   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)