House Prices

library(dplyr)
library(ggplot2)
library(Amelia)

train <- read.csv("~/Desktop/house/train.csv", stringsAsFactors = FALSE)
test <- read.csv("~/Desktop/house/test.csv", stringsAsFactors = FALSE)

missmap(train)

missmap(test)

table(train$PoolQC)

Ex Fa Gd 
 2  2  3 

table(test$PoolQC)

Ex Gd 
 2  1 

table(train$MiscVal)

    0    54   350   400   450   480   500   560   600   620   700   800  1150 
 1408     1     1    11     4     2     8     1     4     1     5     1     1 
 1200  1300  1400  2000  2500  3500  8300 15500 
    2     1     1     4     1     1     1     1 

table(test$MiscVal)

    0    80   300   400   420   450   455   460   490   500   600   650   700 
 1408     1     1     7     1     5     1     1     1     5     4     3     2 
  750   900  1000  1200  1500  1512  2000  2500  3000  4500  6500 12500 17000 
    1     1     1     1     3     1     3     1     2     2     1     1     1 

table(train$Fireplaces)

  0   1   2   3 
690 650 115   5 

table(test$LotFrontage)

 21  22  24  25  26  28  30  31  32  33  34  35  36  37  38  39  40  41  42  43 
 27   1  30   1   3   1   5   1   3   2   6  10   2   2   3   5   4   8   8  11 
 44  45  46  47  48  49  50  51  52  53  54  55  56  57  58  59  60  61  62  63 
 10   7   2   3  10   1  60  12  14  14   4   7  14  14  10  14 133   9  22  30 
 64  65  66  67  68  69  70  71  72  73  74  75  76  77  78  79  80  81  82  83 
 24  49  11  10  25   9  63   7  22  15  24  52  14   6  21  11  68  12  16   8 
 84  85  86  87  88  89  90  91  92  93  94  95  96  97  98  99 100 101 102 103 
  9  36   5   6  12   4  23   8   5   5   6   8   6   3   4   3  12   1   6   3 
104 105 106 107 108 109 110 112 113 114 115 117 118 119 120 121 123 124 125 126 
  2   6   3   4   3   2   9   3   3   3   1   1   2   1   7   1   1   2   3   1 
128 129 130 131 133 134 135 136 140 149 150 155 160 195 200 
  2   2   2   1   1   1   1   1   1   1   1   1   2   1   1 

Garage

train$GarageCond <- ifelse(is.na(train$GarageCond), yes=0, no=train$GarageCond)
train$GarageCond <- ifelse(train$GarageCond == "Ex", yes=5, no=train$GarageCond)
train$GarageCond <- ifelse(train$GarageCond == "Gd", yes=4, no=train$GarageCond)
train$GarageCond <- ifelse(train$GarageCond == "TA", yes=3, no=train$GarageCond)
train$GarageCond <- ifelse(train$GarageCond == "Fa", yes=2, no=train$GarageCond)
train$GarageCond <- ifelse(train$GarageCond == "Po", yes=1, no=train$GarageCond)
test$GarageCond <- ifelse(is.na(test$GarageCond), yes=0, no=test$GarageCond)
test$GarageCond <- ifelse(test$GarageCond == "Ex", yes=5, no=test$GarageCond)
test$GarageCond <- ifelse(test$GarageCond == "Gd", yes=4, no=test$GarageCond)
test$GarageCond <- ifelse(test$GarageCond == "TA", yes=3, no=test$GarageCond)
test$GarageCond <- ifelse(test$GarageCond == "Fa", yes=2, no=test$GarageCond)
test$GarageCond <- ifelse(test$GarageCond == "Po", yes=1, no=test$GarageCond)
train$GarageQual <- ifelse(is.na(train$GarageQual), yes=0, no=train$GarageQual)
train$GarageQual <- ifelse(train$GarageQual == "Ex", yes=5, no=train$GarageQual)
train$GarageQual <- ifelse(train$GarageQual == "Gd", yes=4, no=train$GarageQual)
train$GarageQual <- ifelse(train$GarageQual == "TA", yes=3, no=train$GarageQual)
train$GarageQual <- ifelse(train$GarageQual == "Fa", yes=2, no=train$GarageQual)
train$GarageQual <- ifelse(train$GarageQual == "Po", yes=1, no=train$GarageQual)
test$GarageQual <- ifelse(is.na(test$GarageQual), yes=0, no=test$GarageQual)
test$GarageQual <- ifelse(test$GarageQual == "Ex", yes=5, no=test$GarageQual)
test$GarageQual <- ifelse(test$GarageQual == "Gd", yes=4, no=test$GarageQual)
test$GarageQual <- ifelse(test$GarageQual == "TA", yes=3, no=test$GarageQual)
test$GarageQual <- ifelse(test$GarageQual == "Fa", yes=2, no=test$GarageQual)
test$GarageQual <- ifelse(test$GarageQual == "Po", yes=1, no=test$GarageQual)
train$GarageFinish <- ifelse(is.na(train$GarageFinish), yes=0, no=train$GarageFinish)
train$GarageFinish <- ifelse(train$GarageFinish == "Fin", yes=3, no=train$GarageFinish)
train$GarageFinish <- ifelse(train$GarageFinish == "RFn", yes=2, no=train$GarageFinish)
train$GarageFinish <- ifelse(train$GarageFinish == "Unf", yes=1, no=train$GarageFinish)
test$GarageFinish <- ifelse(is.na(test$GarageFinish), yes=0, no=test$GarageFinish)
test$GarageFinish <- ifelse(test$GarageFinish == "Fin", yes=3, no=test$GarageFinish)
test$GarageFinish <- ifelse(test$GarageFinish == "RFn", yes=2, no=test$GarageFinish)
test$GarageFinish <- ifelse(test$GarageFinish == "Unf", yes=1, no=test$GarageFinish)
test$GarageCars <- ifelse(is.na(test$GarageCars), yes=0, no=test$GarageCars)
garlm <- lm(data = train, formula = GarageArea ~ GarageCars)
pgar <- predict(object = garlm, newdata = test)
test$GarageArea <- ifelse(is.na(test$GarageArea), yes=pgar, no=test$GarageArea)

PoolQC to Numbers

train$PoolQC <- ifelse(is.na(train$PoolQC), yes=0, no=train$PoolQC)
train$PoolQC <- ifelse(train$PoolQC == "Ex", yes=10, no=train$PoolQC)
train$PoolQC <- ifelse(train$PoolQC == "Gd", yes=8, no=train$PoolQC)
train$PoolQC <- ifelse(train$PoolQC == "Fa", yes=3, no=train$PoolQC)
test$PoolQC <- ifelse(is.na(test$PoolQC), yes=0, no=test$PoolQC)
test$PoolQC <- ifelse(test$PoolQC == "Ex", yes=10, no=test$PoolQC)
test$PoolQC <- ifelse(test$PoolQC == "Gd", yes=8, no=test$PoolQC)

Fence to Numbers

train$Fence <- ifelse(is.na(train$Fence), yes=0, no=train$Fence)
train$Fence <- ifelse(train$Fence == "MnWw", yes=1, no=train$Fence)
train$Fence <- ifelse(train$Fence == "GdWo", yes=2, no=train$Fence)
train$Fence <- ifelse(train$Fence == "MnPrv", yes=3, no=train$Fence)
train$Fence <- ifelse(train$Fence == "GdPrv", yes=4, no=train$Fence)
test$Fence <- ifelse(is.na(test$Fence), yes=0, no=test$Fence)
test$Fence <- ifelse(test$Fence == "MnWw", yes=1, no=test$Fence)
test$Fence <- ifelse(test$Fence == "GdWo", yes=2, no=test$Fence)
test$Fence <- ifelse(test$Fence == "MnPrv", yes=3, no=test$Fence)
test$Fence <- ifelse(test$Fence == "GdPrv", yes=4, no=test$Fence)

Alley and Street to Numbers

train$Alley <- ifelse(is.na(train$Alley), yes=0, no=train$Alley)
train$Alley <- ifelse(train$Alley == "Grvl", yes=1, no=train$Alley)
train$Alley <- ifelse(train$Alley == "Pave", yes=2, no=train$Alley)
test$Alley <- ifelse(is.na(test$Alley), yes=0, no=test$Alley)
test$Alley <- ifelse(test$Alley == "Grvl", yes=1, no=test$Alley)
test$Alley <- ifelse(test$Alley == "Pave", yes=2, no=test$Alley)
train$Street <- ifelse(train$Street == "Grvl", yes=1, no=train$Street)
train$Street <- ifelse(train$Street == "Pave", yes=2, no=train$Street)
test$Street <- ifelse(test$Street == "Grvl", yes=1, no=test$Street)
test$Street <- ifelse(test$Street == "Pave", yes=2, no=test$Street)

Fireplace

train$FireplaceQu <- ifelse(is.na(train$FireplaceQu), yes=0, no=train$FireplaceQu)
train$FireplaceQu <- ifelse(train$FireplaceQu == "Ex", yes=5, no=train$FireplaceQu)
train$FireplaceQu <- ifelse(train$FireplaceQu == "Gd", yes=4, no=train$FireplaceQu)
train$FireplaceQu <- ifelse(train$FireplaceQu == "TA", yes=3, no=train$FireplaceQu)
train$FireplaceQu <- ifelse(train$FireplaceQu == "Fa", yes=2, no=train$FireplaceQu)
train$FireplaceQu <- ifelse(train$FireplaceQu == "Po", yes=1, no=train$FireplaceQu)
test$FireplaceQu <- ifelse(is.na(test$FireplaceQu), yes=0, no=test$FireplaceQu)
test$FireplaceQu <- ifelse(test$FireplaceQu == "Ex", yes=5, no=test$FireplaceQu)
test$FireplaceQu <- ifelse(test$FireplaceQu == "Gd", yes=4, no=test$FireplaceQu)
test$FireplaceQu <- ifelse(test$FireplaceQu == "TA", yes=3, no=test$FireplaceQu)
test$FireplaceQu <- ifelse(test$FireplaceQu == "Fa", yes=2, no=test$FireplaceQu)
test$FireplaceQu <- ifelse(test$FireplaceQu == "Po", yes=1, no=test$FireplaceQu)

Varios

train$LotFrontage <- ifelse(is.na(train$LotFrontage), 
                            yes = mean(train$LotFrontage, na.rm = TRUE),
                            no = train$LotFrontage)
test$LotFrontage <- ifelse(is.na(test$LotFrontage), 
                            yes = mean(test$LotFrontage, na.rm = TRUE),
                            no = test$LotFrontage)
test$Utilities <- ifelse(is.na(test$Utilities), yes="AllPub", no=test$Utilities)
test$MSZoning <- ifelse(is.na(test$MSZoning), yes="C", no=test$MSZoning)
test$MSZoning <- ifelse(test$MSZoning == "C (all)", yes="C", no=test$MSZoning)
train$MSZoning <- ifelse(train$MSZoning == "C (all)", yes="C", no=train$MSZoning)
table(test$Utilities)

AllPub 
  1459 

lm <- lm(data = train, formula = SalePrice ~ YrSold + MiscVal + Fence + MSSubClass + MSZoning + LotArea + Utilities + Neighborhood + HouseStyle + OverallQual + OverallCond + YearBuilt + Foundation + GarageArea + PoolArea + MiscVal + SaleCondition + Condition1 + Condition2)
predicted_price <- predict(object = lm, newdata = test)
price <- data.frame(Id = test$Id, SalePrice = round(predicted_price,0))
price$SalePrice <- ifelse (price$SalePrice<0, yes=mean(price$SalePrice), no=price$SalePrice)
write.csv(price, file="prices.csv", row.names = FALSE)