data = read.csv2("https://raw.githubusercontent.com/basilhan/datasets/master/kc-house-data.csv",sep=",",header=T)
library(tidyverse)

glimpse(data)


plot(data$sqft_living, data$price)
reg = lm(price ~ sqft_living, data = data)
abline(reg,col="red")
summary(reg)


library(geosphere)

#Koordinaten zu Zahlen umwandeln
data %>%
  rowwise() %>%
  mutate(lat = as.numeric(lat),
         long = as.numeric(long),
         distanceSeattle = distHaversine(c(long,lat),c(-122.3321,47.5813902))
         ) -> data

#Entfernung berechnen und speichern
#coordinatesSeattle = 47.6062,-122.3321

reg = lm(price ~ sqft_living + sqft_lot + condition + grade + sqft_above + distanceSeattle, data = data)
summary(reg)

data %>%
  mutate(sqft_price = as.numeric(price) / as.numeric(sqft_living)) -> data

reg2 = lm(price ~ sqft_living + waterfront, data = data)
summary(reg2)

#nur Entfernung nach Seattle
reg3 = lm(sqft_price ~ distanceSeattle, data = data)
summary(reg3)

#Entfernung nach Seattle und waterfront
reg3 = lm(sqft_price ~ waterfront + distanceSeattle, data = data)
summary(reg3)

#Entfernung nach Seattle und waterfront und view
reg3 = lm(sqft_price ~ waterfront + view + distanceSeattle, data = data)
summary(reg3)

#Alle unabhängigen Variablen im Datensatz (inkl. Distance) außer Datum, Id, Koordinaten, zipcode, squarefootprice
#Inseratpreis vorhersagen
reg4 = lm(price ~ ., select(data, -date, -id,-long,-lat, -zipcode, - sqft_price))
summary(reg4)