Sale price mean (red line) of data = 180921.2 Ames, Iowa Real Estate Price Predictions
Darwhin Gomez
What factors drive sale price?
Buying or selling a home can be very stressful to experienced buyers and sellers. Sometimes even browsing online can become incredibly frustrating. Two similar homes can vary in price significantly and many times the reasons are not clear.
Lets try to make sense of what factors are actually affecting a home’s price.
Intuition Size, Neighborhood
Neighborhoods
Modeling
As we can see, sale prices vary due to a number of factors. But how can we get a holistic, weighted representation of how these factors influence the sale price? One way is to model it. We can use a regression model to predict house prices and then extract the contributing factors and their impact on the model.
Linear Regression – to understand linear relationships between features and sale price. The model uses 59 predictors.
SalePrice ~ Id + MSSubClass + MSZoning + LotFrontage + LotArea +
LotShape + LandContour + LotConfig + LandSlope + Neighborhood +
Condition1 + Condition2 + BldgType + HouseStyle + OverallQual +
OverallCond + YearBuilt + YearRemodAdd + RoofStyle + RoofMatl +
Exterior1st + Exterior2nd + MasVnrArea + ExterQual + ExterCond +
Foundation + BsmtFinSF1 + BsmtFinSF2 + BsmtUnfSF + TotalBsmtSF +
HeatingQC + Electrical + X1stFlrSF + X2ndFlrSF + LowQualFinSF +
GrLivArea + BsmtFullBath + BsmtHalfBath + FullBath + HalfBath +
BedroomAbvGr + KitchenAbvGr + KitchenQual + TotRmsAbvGrd +
Functional + Fireplaces + GarageYrBlt + GarageCars + GarageArea +
WoodDeckSF + OpenPorchSF + EnclosedPorch + X3SsnPorch + ScreenPorch +
PoolArea + MiscVal + MoSold + YrSold + SaleType + SaleCondition
Call:
lm(formula = SalePrice ~ ., data = train_data)
Residuals:
Min 1Q Median 3Q Max
-105676 -10396 0 9420 148320
Coefficients: (2 not defined because of singularities)
Estimate Std. Error t value Pr(>|t|)
(Intercept) -8.240e+05 1.428e+06 -0.577 0.564059
Id 1.764e+00 2.190e+00 0.805 0.420971
MSSubClass 8.530e+01 1.319e+02 0.647 0.517934
MSZoningFV 2.839e+04 1.553e+04 1.828 0.067966 .
MSZoningRH 3.936e+04 1.623e+04 2.425 0.015584 *
MSZoningRL 2.883e+04 1.386e+04 2.080 0.037895 *
MSZoningRM 2.412e+04 1.293e+04 1.865 0.062592 .
LotFrontage 2.717e+01 6.131e+01 0.443 0.657821
LotArea 1.729e+00 3.167e-01 5.459 6.76e-08 ***
LotShapeIR2 -1.751e+03 5.983e+03 -0.293 0.769882
LotShapeIR3 6.487e+03 1.289e+04 0.503 0.615012
LotShapeReg -1.317e+03 2.360e+03 -0.558 0.577020
LandContourHLS 1.378e+04 7.278e+03 1.894 0.058656 .
LandContourLow -1.599e+04 1.071e+04 -1.493 0.135920
LandContourLvl 3.483e+03 5.617e+03 0.620 0.535380
LotConfigCulDSac 9.435e+03 5.666e+03 1.665 0.096332 .
LotConfigFR2 -7.637e+03 6.165e+03 -1.239 0.215858
LotConfigFR3 -9.755e+03 1.437e+04 -0.679 0.497583
LotConfigInside -1.985e+02 2.556e+03 -0.078 0.938132
LandSlopeMod 9.081e+03 5.356e+03 1.695 0.090453 .
LandSlopeSev -8.330e+04 2.235e+04 -3.727 0.000211 ***
NeighborhoodBlueste 9.184e+03 2.023e+04 0.454 0.649944
NeighborhoodBrDale 1.924e+04 1.386e+04 1.388 0.165615
NeighborhoodBrkSide 1.482e+04 1.277e+04 1.160 0.246264
NeighborhoodClearCr -1.096e+04 1.398e+04 -0.784 0.433189
NeighborhoodCollgCr -5.723e+03 9.110e+03 -0.628 0.530031
NeighborhoodCrawfor 1.603e+04 1.109e+04 1.445 0.148799
NeighborhoodEdwards -9.072e+03 1.017e+04 -0.892 0.372777
NeighborhoodGilbert -1.399e+04 1.020e+04 -1.372 0.170646
NeighborhoodIDOTRR 1.195e+04 1.527e+04 0.783 0.434179
NeighborhoodMeadowV 1.563e+04 1.550e+04 1.008 0.313660
NeighborhoodMitchel -1.567e+04 1.083e+04 -1.447 0.148320
NeighborhoodNAmes -7.468e+03 1.000e+04 -0.747 0.455455
NeighborhoodNoRidge 1.041e+04 1.090e+04 0.956 0.339592
NeighborhoodNPkVill 5.318e+03 2.537e+04 0.210 0.834026
NeighborhoodNridgHt 1.468e+04 8.913e+03 1.647 0.100054
NeighborhoodNWAmes -2.088e+04 1.067e+04 -1.957 0.050819 .
NeighborhoodOldTown 4.900e+02 1.284e+04 0.038 0.969581
NeighborhoodSawyer -1.225e+03 1.071e+04 -0.114 0.908973
NeighborhoodSawyerW -1.375e+03 1.008e+04 -0.136 0.891548
NeighborhoodSomerst 9.827e+03 1.065e+04 0.923 0.356577
NeighborhoodStoneBr 4.482e+04 1.100e+04 4.074 5.18e-05 ***
NeighborhoodSWISU 2.063e+03 1.257e+04 0.164 0.869668
NeighborhoodTimber -1.317e+04 1.076e+04 -1.224 0.221347
NeighborhoodVeenker -5.384e+03 1.582e+04 -0.340 0.733669
Condition1Feedr -4.124e+03 6.728e+03 -0.613 0.540092
Condition1Norm 9.227e+03 5.181e+03 1.781 0.075381 .
Condition1PosA 4.317e+03 1.679e+04 0.257 0.797191
Condition1PosN -1.074e+03 1.299e+04 -0.083 0.934117
Condition1RRAe -1.493e+04 1.133e+04 -1.318 0.187929
Condition1RRAn 1.655e+03 8.745e+03 0.189 0.849965
Condition1RRNe 2.747e+03 2.477e+04 0.111 0.911749
Condition1RRNn -1.618e+04 1.792e+04 -0.903 0.367006
Condition2Feedr 8.159e+02 3.453e+04 0.024 0.981157
Condition2Norm 9.581e+03 2.937e+04 0.326 0.744317
Condition2PosA 2.755e+04 5.039e+04 0.547 0.584837
Condition2PosN -4.649e+04 4.059e+04 -1.146 0.252386
BldgType2fmCon -1.354e+04 1.856e+04 -0.730 0.465924
BldgTypeDuplex -1.315e+04 1.186e+04 -1.109 0.267843
BldgTypeTwnhs -3.348e+04 1.516e+04 -2.208 0.027579 *
BldgTypeTwnhsE -2.671e+04 1.430e+04 -1.867 0.062342 .
HouseStyle1.5Unf 1.455e+04 1.231e+04 1.182 0.237642
HouseStyle1Story 1.676e+04 6.122e+03 2.738 0.006340 **
HouseStyle2.5Fin -4.227e+04 1.718e+04 -2.461 0.014114 *
HouseStyle2.5Unf -2.614e+04 1.260e+04 -2.075 0.038373 *
HouseStyle2Story -3.083e+03 4.900e+03 -0.629 0.529513
HouseStyleSFoyer 1.067e+04 8.620e+03 1.238 0.216099
HouseStyleSLvl 6.578e+03 7.877e+03 0.835 0.403975
OverallQual 8.506e+03 1.435e+03 5.929 4.91e-09 ***
OverallCond 7.081e+03 1.261e+03 5.615 2.89e-08 ***
YearBuilt 3.417e+02 1.135e+02 3.010 0.002709 **
YearRemodAdd 1.145e+02 8.147e+01 1.406 0.160196
RoofStyleGable 4.026e+04 2.845e+04 1.415 0.157558
RoofStyleGambrel 3.771e+04 3.036e+04 1.242 0.214613
RoofStyleHip 3.670e+04 2.851e+04 1.287 0.198525
RoofStyleMansard 5.427e+04 3.210e+04 1.691 0.091354 .
RoofMatlMembran 1.584e+05 4.373e+04 3.621 0.000316 ***
RoofMatlRoll -2.566e+03 2.726e+04 -0.094 0.925037
RoofMatlTar&Grv 2.424e+04 2.521e+04 0.961 0.336784
RoofMatlWdShake -2.398e+04 2.566e+04 -0.934 0.350402
RoofMatlWdShngl 6.379e+04 1.324e+04 4.818 1.80e-06 ***
Exterior1stBrkComm -2.767e+04 3.693e+04 -0.749 0.454039
Exterior1stBrkFace -4.554e+02 1.509e+04 -0.030 0.975933
Exterior1stCemntBd -4.994e+04 3.409e+04 -1.465 0.143399
Exterior1stHdBoard -1.265e+04 1.571e+04 -0.806 0.420705
Exterior1stImStucc -3.893e+04 3.094e+04 -1.258 0.208765
Exterior1stMetalSd -5.856e+03 1.814e+04 -0.323 0.746941
Exterior1stPlywood -1.726e+04 1.596e+04 -1.082 0.279762
Exterior1stStone -5.172e+04 3.335e+04 -1.551 0.121431
Exterior1stStucco -1.783e+04 1.682e+04 -1.060 0.289367
Exterior1stVinylSd -1.533e+04 1.605e+04 -0.955 0.339923
Exterior1stWd Sdng -1.997e+04 1.496e+04 -1.336 0.182156
Exterior1stWdShing -5.319e+03 1.590e+04 -0.334 0.738133
Exterior2ndAsphShn 1.221e+04 2.553e+04 0.478 0.632604
Exterior2ndBrk Cmn 3.019e+04 3.045e+04 0.991 0.321833
Exterior2ndBrkFace 7.428e+03 1.552e+04 0.479 0.632337
Exterior2ndCmentBd 4.638e+04 3.375e+04 1.374 0.169857
Exterior2ndHdBoard 8.093e+03 1.505e+04 0.538 0.590819
Exterior2ndImStucc 1.431e+04 1.833e+04 0.781 0.435251
Exterior2ndMetalSd 7.156e+03 1.753e+04 0.408 0.683255
Exterior2ndOther -2.519e+04 2.964e+04 -0.850 0.395641
Exterior2ndPlywood 5.921e+03 1.451e+04 0.408 0.683245
Exterior2ndStone 2.126e+04 2.077e+04 1.024 0.306376
Exterior2ndStucco 9.946e+03 1.564e+04 0.636 0.525152
Exterior2ndVinylSd 1.351e+04 1.547e+04 0.873 0.383006
Exterior2ndWd Sdng 1.763e+04 1.403e+04 1.256 0.209495
Exterior2ndWd Shng 4.407e+03 1.474e+04 0.299 0.764975
MasVnrArea 2.841e+01 6.784e+00 4.187 3.20e-05 ***
ExterQualFa -1.263e+04 1.630e+04 -0.775 0.438620
ExterQualGd -3.465e+04 6.321e+03 -5.482 5.97e-08 ***
ExterQualTA -3.423e+04 7.282e+03 -4.701 3.14e-06 ***
ExterCondFa 2.197e+03 2.724e+04 0.081 0.935751
ExterCondGd -3.417e+03 2.622e+04 -0.130 0.896371
ExterCondTA -1.945e+03 2.617e+04 -0.074 0.940777
FoundationCBlock 4.822e+03 4.387e+03 1.099 0.272161
FoundationPConc 5.646e+03 4.812e+03 1.173 0.241083
FoundationSlab 1.978e+04 1.044e+04 1.895 0.058525 .
FoundationStone 1.886e+04 1.622e+04 1.163 0.245115
FoundationWood -3.894e+04 1.969e+04 -1.977 0.048404 *
BsmtFinSF1 4.176e+01 5.806e+00 7.191 1.73e-12 ***
BsmtFinSF2 2.947e+01 7.478e+00 3.941 8.99e-05 ***
BsmtUnfSF 1.655e+01 5.417e+00 3.055 0.002337 **
TotalBsmtSF NA NA NA NA
HeatingQCFa 1.948e+03 7.145e+03 0.273 0.785237
HeatingQCGd -3.171e+03 2.942e+03 -1.078 0.281450
HeatingQCPo 1.353e+03 3.038e+04 0.045 0.964483
HeatingQCTA -4.524e+03 2.893e+03 -1.563 0.118418
ElectricalFuseF 8.728e+03 7.961e+03 1.096 0.273337
ElectricalFuseP -4.012e+04 3.396e+04 -1.181 0.237971
ElectricalMix 2.685e+04 3.116e+04 0.862 0.389195
ElectricalSBrkr 8.846e+02 4.110e+03 0.215 0.829637
X1stFlrSF 6.098e+01 7.631e+00 7.992 5.90e-15 ***
X2ndFlrSF 7.893e+01 7.313e+00 10.793 < 2e-16 ***
LowQualFinSF 8.852e+01 3.087e+01 2.867 0.004273 **
GrLivArea NA NA NA NA
BsmtFullBath 1.619e+03 2.613e+03 0.620 0.535753
BsmtHalfBath 2.730e+03 4.186e+03 0.652 0.514562
FullBath 3.116e+03 3.083e+03 1.011 0.312506
HalfBath -4.122e+03 2.981e+03 -1.383 0.167171
BedroomAbvGr -7.461e+03 1.882e+03 -3.965 8.15e-05 ***
KitchenAbvGr -1.464e+04 9.035e+03 -1.621 0.105516
KitchenQualFa -1.180e+04 9.021e+03 -1.308 0.191447
KitchenQualGd -2.335e+04 4.338e+03 -5.383 1.02e-07 ***
KitchenQualTA -2.108e+04 5.185e+03 -4.066 5.35e-05 ***
TotRmsAbvGrd 1.004e+03 1.303e+03 0.770 0.441415
FunctionalMaj2 -2.736e+04 1.904e+04 -1.437 0.151072
FunctionalMin1 -1.228e+04 1.209e+04 -1.016 0.309791
FunctionalMin2 -5.099e+03 1.211e+04 -0.421 0.673815
FunctionalMod -2.140e+04 1.359e+04 -1.574 0.115912
FunctionalTyp 2.317e+03 1.035e+04 0.224 0.823026
Fireplaces 1.941e+03 1.944e+03 0.998 0.318578
GarageYrBlt 1.366e+02 8.379e+01 1.630 0.103613
GarageCars 6.132e+03 2.956e+03 2.074 0.038458 *
GarageArea 2.617e+00 1.043e+01 0.251 0.801894
WoodDeckSF 1.022e+01 8.339e+00 1.226 0.220691
OpenPorchSF 2.826e+01 1.681e+01 1.681 0.093172 .
EnclosedPorch 1.559e+01 1.807e+01 0.862 0.388750
X3SsnPorch 4.304e+01 3.568e+01 1.206 0.228106
ScreenPorch 3.910e+01 1.665e+01 2.349 0.019130 *
PoolArea -5.012e+00 3.254e+01 -0.154 0.877651
MiscVal -3.475e+00 6.273e+00 -0.554 0.579795
MoSold -9.337e+02 3.407e+02 -2.740 0.006303 **
YrSold -2.188e+02 7.105e+02 -0.308 0.758223
SaleTypeCon 4.834e+04 2.811e+04 1.719 0.086020 .
SaleTypeConLD -1.054e+04 1.364e+04 -0.773 0.439852
SaleTypeConLI 5.644e+02 1.593e+04 0.035 0.971753
SaleTypeConLw 3.163e+03 1.574e+04 0.201 0.840788
SaleTypeCWD 1.508e+04 1.582e+04 0.953 0.340900
SaleTypeNew -2.561e+03 1.828e+04 -0.140 0.888598
SaleTypeOth 3.292e+04 2.444e+04 1.347 0.178466
SaleTypeWD -5.162e+03 5.883e+03 -0.877 0.380540
SaleConditionAdjLand 2.946e+04 2.630e+04 1.120 0.263008
SaleConditionAlloca 9.006e+03 1.191e+04 0.756 0.449964
SaleConditionFamily 4.548e+03 8.073e+03 0.563 0.573391
SaleConditionNormal 7.142e+03 4.138e+03 1.726 0.084859 .
SaleConditionPartial 2.156e+04 1.744e+04 1.236 0.216830
---
Signif. codes: 0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
Residual standard error: 23240 on 664 degrees of freedom
(259 observations deleted due to missingness)
Multiple R-squared: 0.9406, Adjusted R-squared: 0.9251
F-statistic: 60.73 on 173 and 664 DF, p-value: < 2.2e-16[1] " Train Rsquared"[1] 0.9405563[1] "Test Rsquared"[1] 0.7471636[1] "Test RMSE"[1] 38826.67Extracted Significant Features
Model Residuals
Predictions
Predictions Other Significant Features
The Roof
Takeaways
Price is directly impacted by square footage, homes with more livable square footage tend to sell higher.
Price can vary significantly by neighborhood.
Better overall quality homes sell at higher prices even for comparable Lot sizes.
Building Materials for the roof have a significant impact on sale price.
Data Code
** The Model trained here did not incorporate all original predictors, predictors with non factor levels, and exactly correlated predictors like GrLivArea. Also visualizations show model predictions on 75/25 split not the entire data set.
Thanks