twitter.Rmd

---
title: "R Notebook"
output: html_notebook
editor_options: 
  chunk_output_type: inline
---

# Gelistirici Hesabi Acilmasi


# 1. API Baglantisi ve Paket Kurulumlari

```{r}
library(twitteR)
library(ROAuth)
library(tm)
library(RCurl) 
options(RCurlOptions = list(cainfo = system.file("CurlSSL", "cacert.pem", package = "RCurl")))


#TWEETTER DAN ELDE EDILMIS KODLAR
api_key <- "2IaJWH6sdjqhY0JGJvdSTKkVk"
api_secret <- "lt1BjcuDbmwZxNnh90gK2tdaq9gWv87M0DVLmwYc0ZX0RPZGHb"
access_token <- "2951805075-k7CWKll7vWeIDJZeeX5LEtUGKtYO0okU8UwdlyU"
access_token_secret <- "4FNoJj06RItB1VnZbF5HbcGRaVvRNPgxf5HGQKOvqaMNG"


# registerTwitterOAuth(twitCred)
setup_twitter_oauth(api_key, api_secret, access_token, access_token_secret)

#updateStatus("Bu tweet udemy egitimindeki Twitter analitigi bolumu icin atilmistir.")

```







# 2. Veri Cekme

## Hashtag'lerden Veri Cekmek

```{r}
library(tidyverse)
availableTrendLocations() %>% filter(country == "Turkey")

h <- getTrends(2344116)
head(h,20)
tw <- searchTwitter("#40BinÖğrtAtamasıHaktır", n = 2000)


class(tw)
str(tw)

df_tw <- twListToDF(tw)
View(df_tw)
```

## Profillerden Veri Cekmek

```{r}
df_user <- userTimeline('mvahitkeskin', n = 100)
df <- twListToDF(df_user)
View(df)

```
























# 3.Profil Analizi


## Temel Bilgilerin Cekilmesi
```{r}

mvk <- getUser("mvahitkeskin")
attributes(mvk)
str(mvk)
mvk$name
mvk$id
mvk$screenName
mvk$created
mvk$url
mvk$location
mvk$statusesCount
mvk$followersCount
mvk$favoritesCount
mvk$friendsCount
mvk$profileImageUrl
download.file("http://pbs.twimg.com/profile_images/855403183266570244/Dk1UJdMW_normal.jpg", 
              destfile = "pl.jpg")

mvk$getFavorites(n=3)

mvk$getFriends(n=10)


atom <- getUser("GuveliTugbanur")
atom$favoritesCount

mvk$getFollowerIDs(n = 100)

mvk$lastStatus$statusSource

```























## Profilin En'leri 

```{r}
df_user <- userTimeline('fatihportakal', n = 2000)

df <- twListToDF(df_user)

df %>% 
  select(text, favoriteCount) %>%
  arrange(desc(favoriteCount)) %>%
  top_n(5) %>%
  View()


df %>% 
  select(text, retweetCount) %>%
  arrange(desc(retweetCount)) %>%
  top_n(5) %>%
  View()


```

## Retweet ve Favori Dagilimlari
```{r}
df_user <- userTimeline('fatihportakal', n = 2000)
df <- twListToDF(df_user)
c <- data.frame(fav = df$favoriteCount, ret = df$retweetCount)
ggplot(data = c, aes(fav)) + geom_density()
library(funModeling)
profiling_num(c)

ggplot(c,aes(fav)) + 
  geom_histogram(aes(y=..density..), colour = "black", fill = "white") +
  geom_density(alpha = 0.3, fill = "orange")

```
























## Kullanim Saatleri Dagilimi

```{r}
df_user <- userTimeline('fatihportakal', n = 2000)
df <- twListToDF(df_user)
library(lubridate)

hist(hour(df$created), col = "purple", 
     xlab = "Saat Araligi", 
     ylab = "Tweet Sayisi",
     xlim = c(0,25))

gunisim <- wday(df$created, label = TRUE)

ggplot(df, aes(gunisim)) + geom_bar()



```















## Baglanma Kaynaklari

```{r}
df_user <- userTimeline("fatihportakal", n=5000) 
df <- twListToDF(df_user)
df$statusSource[1]

kaynaklar <- df$statusSource

kaynaklar <- gsub("</a>","", kaynaklar)
kaynaklar <- strsplit(kaynaklar, ">")
kaynaklar <- sapply(kaynaklar, function(x) x[2])
kaynak_tablosu <- table(kaynaklar)
pie(kaynak_tablosu, radius = 0.9, border = 8)
```























##  Takipcilerin Analizi

```{r}

v <- getUser("mvahitkeskin")
takipciler <- v$getFollowers()
df <- twListToDF(takipciler)
View(df)


df %>% 
  mutate(popi_indeksi = friendsCount / followersCount) %>%
  select(name, description, popi_indeksi) %>%
  arrange(desc(popi_indeksi)) %>%
  top_n(5)



df %>% 
  filter(followersCount > friendsCount) %>%
  mutate(popi_indeksi = friendsCount / followersCount) %>%
  select(name, description, popi_indeksi) %>%
  arrange(desc(popi_indeksi)) %>%
  top_n(5)

df %>% 
  filter(followersCount > friendsCount) %>%
  mutate(farklar = followersCount - friendsCount) %>%
  select(farklar) %>%
  summarise(n = n(),
            mean = mean(farklar),
            median = mean(farklar),
            sd = sd(farklar))


df %>% 
  filter(followersCount > friendsCount) %>%
  mutate(farklar = followersCount - friendsCount) %>%
  filter(farklar > 1000) %>%
  mutate(popi_indeksi = friendsCount / followersCount) %>%
  select(name, description, popi_indeksi) %>%
  arrange(desc(popi_indeksi)) %>%
  top_n(5)

```


## Takip Edilenlerin Analizi

```{r}

arkadaslar <- v$getFriends()
df <- twListToDF(arkadaslar)
df$location


ggplot(df, aes(df$location)) + geom_bar()

df$location <- sapply(df$location, function(x) ifelse(nchar(x) > 0, x, NA ))

df <- df[!is.na(df$location),]


ggplot(df, aes(location) ) + geom_bar()


a <- df %>% group_by(location) %>%
  summarise(n = n())

b <- a %>% filter(n > mean(a$n))

ggplot(b, aes(b$location) ) + geom_bar()

```
















# 4.Hashtag Analizi


## Trendlere Erismek
```{r}

availableTrendLocations() %>% filter(country == "Turkey")

getTrends(woeid = 23424969)

a <- searchTwitter("#datascience", n = 2000)
df<- twListToDF(a)
View(df)

```

















## Hashtag Betimleme

### Etikete katilim saglayan essiz kac kisi var?
```{r}

df %>% distinct(screenName) %>% count()


```

















### Etikete en cok katki saglayan 5 kisi kimdir?

```{r}
df %>% group_by(screenName) %>%
  summarise(n = n()) %>%
  arrange(desc(n)) %>%
  top_n(20)

```
















### Etikete Katma Deger Saglayan En Degerli 5 kisi Kimdir?


```{r}
df %>% 
  filter(isRetweet == "FALSE") %>%
  group_by(screenName) %>%
  summarise(n = n()) %>%
  arrange(desc(n)) %>%
  top_n(10)


updateStatus("You are the best original contributer for this hashtag #datascience. 
             @IainLJBrown-23, programmingncr-19, BristowColin-14")

mvk <- getUser("mvahitkeskin")
mvk$lastStatus$text


searchTwitter("#datascience", n = 100, resultType = "recent")

```


















### En cok favlanan 5 twit
```{r}

df %>% select(text, screenName, favoriteCount) %>%
  arrange(desc(favoriteCount)) %>%
  top_n(5) %>% View()

```












### En cok retweet edilen 5 twit

```{r}


df %>% select(text, screenName, statusSource, retweetCount) %>%
  arrange(desc(retweetCount)) %>%
  top_n(50) %>% View()

```



















### Tweet Saat Dagilimi

```{r}

a <- searchTwitter("#datascience", n = 5000)
df <- twListToDF(a)

library(lubridate)

hist(hour(df$created), col = "purple", xlim = c(5,24))

```
















### Kaynak Dagilimi

```{r}


df$statusSource[1]

kaynaklar <- df$statusSource

kaynaklar <- gsub("</a>","", kaynaklar)
kaynaklar <- strsplit(kaynaklar, ">")
kaynaklar <- sapply(kaynaklar, function(x) x[2])
kaynak_tablosu <- table(kaynaklar)
pie(kaynak_tablosu, radius = 0.9, border = 8)

df <- data.frame(kaynak_tablosu) 
df <- df %>% filter(Freq >50)


ggplot(df, aes(kaynaklar, Freq)) + geom_bar(stat = "identity") 



```





























# 5.UYGULAMALAR 

## UYGULAMA I - ipad mi iphone mu?
```{r}
df_user <- userTimeline("fatihportakal", n=2000) 
df <- twListToDF(df_user)
View(df)

kaynaklar <- df$statusSource
kaynaklar <- gsub("</a>","", kaynaklar)
kaynaklar <- strsplit(kaynaklar, ">")
kaynaklar <- sapply(kaynaklar, function(x) x[2])
kaynak_tablosu <- table(kaynaklar)


nrow(df)
length(kaynaklar)
df$kaynaklar <- kaynaklar

test_df <- df %>% filter(kaynaklar == "Twitter for iPad" | 
                           kaynaklar == "Twitter for iPhone") %>%
  select(kaynaklar, retweetCount, favoriteCount)



t.test(retweetCount ~ kaynaklar, data = test_df)



```




















## UYGULAMA II - Twitter Metin Madencilgi Kendinizi 5 Kelime ile Anlatabilir Misiniz?
```{r}
df_user <- userTimeline('fatihportakal', n = 2000)
df <- twListToDF(df_user)


doc.corpus <- Corpus(VectorSource(df$text))
doc.corpus <- tm_map(doc.corpus, content_transformer(tolower))
doc.corpus <- tm_map(doc.corpus, content_transformer(removePunctuation))
doc.corpus <- tm_map(doc.corpus,content_transformer(removeNumbers))

removeURL <- function(x) gsub("http[[:alnum:]]*", "", x)
myCorpus <- tm_map(doc.corpus, removeURL)
myCorpus <- tm_map(myCorpus, stripWhitespace)#bosluklarin
tdm <- TermDocumentMatrix(myCorpus)
findFreqTerms(tdm, lowfreq = 5)









#YABANCI

df_user <- userTimeline('AndrewYNg', n = 2000)
df <- twListToDF(df_user)

doc.corpus <- Corpus(VectorSource(df$text))
doc.corpus <- tm_map(doc.corpus, content_transformer(tolower))
doc.corpus <- tm_map(doc.corpus, content_transformer(removePunctuation))
doc.corpus <- tm_map(doc.corpus,content_transformer(removeNumbers))
doc.corpus <- tm_map(doc.corpus, removeWords, stopwords("english"))
removeURL <- function(x) gsub("http[[:alnum:]]*", "", x)
myCorpus <- tm_map(doc.corpus, removeURL)
library(SnowballC)
doc.corpus <- tm_map(doc.corpus, stemDocument)
myCorpus <- tm_map(myCorpus, stripWhitespace)#bosluklarin temizlenmesi

inspect(myCorpus[11:15])

tdm <- TermDocumentMatrix(myCorpus)

findFreqTerms(tdm, lowfreq = 20)

```













## UYGULAMA III - Arkadasini Soyle Kim Oldugunu Soyleyeyim
```{r}
v <- getUser("jtleek")
arkadaslar <- v$getFriends()
df_jt <- twListToDF(arkadaslar)
doc.corpus <- Corpus(VectorSource(df_jt$description))
doc.corpus <- tm_map(doc.corpus, content_transformer(tolower))
doc.corpus <- tm_map(doc.corpus, content_transformer(removePunctuation))
doc.corpus <- tm_map(doc.corpus,content_transformer(removeNumbers))
doc.corpus <- tm_map(doc.corpus, removeWords, stopwords("english"))
removeURL <- function(x) gsub("http[[:alnum:]]*", "", x)
myCorpus <- tm_map(doc.corpus, removeURL)
library(SnowballC)
doc.corpus <- tm_map(doc.corpus, stemDocument)
myCorpus <- tm_map(myCorpus, stripWhitespace)#bosluklarin temizlenmesi

View(df_jt)

tdm <- TermDocumentMatrix(myCorpus)

findFreqTerms(tdm, lowfreq = 40)


```