motorcycler:

# erste Analysen -----
library(dplyr)
library(ggplot2)
library(GGally)
gesammelte_daten <- readr::read_rds("data-chunks/2016-08-28.rds")
with(gesammelte_daten, cor(preis, ez, use = "pairwise.complete.obs"))

hornet_stats %>%
  as.data.frame() %>%
  ggscatmat() +
  geom_smooth()


ggplot(gesammelte_daten, aes(ez, preis)) +
  geom_jitter(width = .4, height = .1) +
  geom_smooth(span = .8) +
  facet_wrap(~motorrad, scales = "free")

# explore combined data ------
# use joins to find data that changed?
library(ggplot2)
library(dplyr)
combined_data <- readr::read_rds("data-combined/combined_data.rds")

# gab es veränderungen im Preis?
ggplot(combined_data, aes(fetched, preis, group = id, colour = motorrad)) +
  geom_line()

combined_data %>%
  group_by(id) %>%
  mutate(changed = case_when(unique(.$preis) > 1 ~ "yes",
                             TRUE                ~ "no"))

augmented <- combined_data %>%
  group_by(id) %>%
  mutate(preis_min = min(preis),
         preis_max = max(preis),
         preis_changed = !identical(preis_min, preis_max)) %>%
  ungroup()

augmented %>%
  filter(preis_changed) %>%
  ggplot(aes(fetched, preis, group = id, colour = id)) +
  geom_line()

# Aktuellster Datensatz: augmented.rds ----
library(ggplot2)
library(dplyr)
library(directlabels)
augmented <- readr::read_rds("data-combined/augmented.rds")

augmented %>%
  filter(preis_changed) %>%
  ggplot(aes(fetched, preis)) +
  geom_line(aes(group = id, colour = ez)) +
  stat_summary(fun.y = "median", geom = "line", colour = "red3") +
  facet_wrap(~motorrad)


# Entwicklung der Preise über die Zeit
p <- ggplot(augmented, aes(fetched, preis, group = motorrad, colour = motorrad)) +
  stat_summary(fun.y = "median", geom = "line")
direct.label(p, list(last.points, hjust = .88))


ggplot(augmented, aes(fetched, preis)) +
  stat_summary(fun.y = "median", geom = "line") +
  facet_wrap(~motorrad)

augmented %>%
  select(-fetched) %>%
  distinct() %>%
  ggplot(aes(ez, preis)) +
  geom_jitter() +
  facet_wrap(~motorrad) +
  geom_smooth()


# how is number of motorcycles available at market changing?
augmented %>%
  group_by(fetched) %>%
  count(motorrad) %>%
  ggplot(aes(fetched, n, colour = motorrad)) +
  geom_line()

# There is something wrong with how many pages are scraped!!!
# Now it is fixed...