Enalquiler

Vectorized extraction from HTML

estatedf <- read.table(file = "/home/xps13/src/R/estate/inst/extdata/enalquiler-2017-10-04.tsv", header = TRUE, sep = "\t")

data_plot <-
  estatedf %>%
  filter(price < 1000 &
         bath < 3)

ggplot(data = data_plot,  aes(x = price, y = size, color = factor(bath))) +
  geom_point() +
  geom_smooth(method = "lm") +
  ggtitle("size: m^2, price: euro") +

  theme(
    legend.position = "top"
  )

str(estatedf)

## 'data.frame':    841 obs. of  6 variables:
##  $ price   : int  650 750 1300 900 2500 750 700 1400 775 500 ...
##  $ size    : int  75 90 157 125 150 65 90 160 75 80 ...
##  $ bath    : int  1 2 2 2 2 2 1 2 1 1 ...
##  $ title   : Factor w/ 556 levels "88 m2, nuevo, zona San Lorenzo Alameda, amueblado, 2 hab",..: 412 205 70 65 57 470 457 322 412 322 ...
##  $ location: Factor w/ 499 levels "Adelantado, Macarena, Resolana, Sevilla, Sevilla",..: 484 172 480 416 85 112 460 334 485 135 ...
##  $ link    : Factor w/ 841 levels "http://www.enalquiler.com/alquiler_piso_sevilla/88-m2-nuevo-zona-san-lorenzo-alameda-amueblado-2-hab_3324700.html",..: 681 346 178 164 130 466 437 521 680 520 ...

mod1 <- lm(price ~ size * bath, data = data_plot)
summary(mod1)

## 
## Call:
## lm(formula = price ~ size * bath, data = data_plot)
## 
## Residuals:
##     Min      1Q  Median      3Q     Max 
## -311.54  -93.96   -0.02   89.22  313.99 
## 
## Coefficients:
##             Estimate Std. Error t value Pr(>|t|)    
## (Intercept) 223.2962    66.0363   3.381 0.000762 ***
## size          5.0343     0.7936   6.343  4.1e-10 ***
## bath        146.7953    50.9886   2.879 0.004115 ** 
## size:bath    -1.4358     0.5489  -2.616 0.009099 ** 
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Residual standard error: 124.7 on 679 degrees of freedom
## Multiple R-squared:  0.2574, Adjusted R-squared:  0.2541 
## F-statistic: 78.45 on 3 and 679 DF,  p-value: < 2.2e-16

knitr::kable(head(estatedf))

price	size	bath	title	location	link
650	75	1	Piso en Sevilla	Triana, Triana - Esperanza de Triana, Sevilla, Sevilla	http://www.enalquiler.com/alquiler_piso_sevilla/piso-en-sevilla_4848337.html
750	90	2	Estupendo piso de alquiler sin amueblar en nervión	Diego Martinez Barrio, Nervión, Diego Martinez Barrio, San Bernardo, Sevilla, Sevilla	http://www.enalquiler.com/alquiler_piso_sevilla/estupendo-piso-de-alquiler-sin-amueblar-en-nervion_4848170.html
1300	157	2	Alquiler piso ascensor y aire acondicionado Triana	Triana, Pagés del Corro-López de Gomara, Triana Casco Antiguo, Sevilla, Sevilla	http://www.enalquiler.com/alquiler_piso_sevilla/alquiler-piso-ascensor-y-aire-acondicionado-triana_4847776.html
900	125	2	Alquiler piso ascensor San pablo - santa justa	San Pablo - Santa Justa, Luis Montoto - Santa Justa, San José Obrero, Sevilla, Sevilla	http://www.enalquiler.com/alquiler_piso_sevilla/alquiler-piso-ascensor-san-pablo-santa-justa_4847728.html
2500	150	2	Alquiler piso ascensor Casco antiguo	Casco Antiguo, Arenal - Museo, Museo, Sevilla, Sevilla	http://www.enalquiler.com/alquiler_piso_sevilla/alquiler-piso-ascensor-casco-antiguo_4847725.html
750	65	2	Piso, Centro - Puerta Osario, 2 dormitorios.	Casco Antiguo, Puerta Osario, Santa Catalina, Sevilla, Sevilla	http://www.enalquiler.com/alquiler_piso_sevilla/piso-centro-puerta-osario-2-dormitorios_4847541.html

The Queen

Vectorized extraction from HTML

Contents