gofund.Rmd

---
title: "R Notebook"
output:
  html_document:
    df_print: paged
---

#Importing libraries
```{r}
library(tidyverse)
library(ggplot2)
library(stringi)
library(wordcloud)
library(tidytext)
library(stringr)
library(dplyr)
library(textdata)
```

```{r}
#importing the dataset
gofund <- read.csv("gofundme_data_OSF.csv")
head(gofund)
```


```{r}
#Preprocessing
#Encoding categorical data
gofund$currency <- factor(gofund$currency, levels = c("USD", "GBP", "CAD","AUD","EUR","NOK"))
gofund$Category <- factor(gofund$Category, levels = c("Funerals & Memorials","Sports, Teams & Clubs","Medical, Illness & Healing", "Missions, Faith & Church" ,"Business & Entrepreneurs","Babies, Kids & Family","Other","Education & Learning", "Community & Neighbors","Accidents & Emergencies","Creative Arts, Music & Film","Celebrations & Events","Volunteer & Service", "Animals & Pets" ,  "Rent, Food & Monthly Bills",  "Environment", "Non-Profits & Charities",  "Dreams, Hopes & Wishes",  "Competitions & Pageants","Travel & Adventure"))  

#Success ==  1 | Fail == 0
gofund$state <- ifelse(gofund$Raised > gofund$Goal,1,0)
```

```{r}
#Data cleaning
gofund<- na.omit(gofund)


gofund$negate <- as.numeric(gofund$negate)
gofund$sad <- as.numeric(gofund$sad)
gofund$risk <- as.numeric(gofund$risk)
gofund$money <- as.numeric(gofund$money)
gofund$anger <- as.numeric(gofund$anger)
gofund$posemo <- as.numeric(gofund$posemo)
gofund$negemo <- as.numeric(gofund$negemo)
gofund$anx <- as.numeric(gofund$anx)
gofund$anger <- as.numeric(gofund$anger)
gofund$sad <- as.numeric(gofund$sad)
gofund$risk <- as.numeric(gofund$risk)
gofund$money <- as.numeric(gofund$money)

gofund<- na.omit(gofund)
```

```{r}
#Data Exploration
#Failed vs Success
gofund %>% group_by(state) %>% count(state)
```
```{r}
#Number of countries
levels(gofund$currency)

```
```{r}
#success/failure rate  country
grp_by_country <- gofund %>%
  group_by(currency) %>%
  summarise(successful = sum(state == 1), failed = sum(state == 0)) 
grp_by_country
```
```{r}
#Visualization
ggplot(gofund , aes(currency)) + geom_bar(aes(fill=state))
```
```{r}
#success/failure rate category
grp_by_category <- gofund %>%
  group_by(Category) %>%
  summarise(successful = sum(state == 1), failed = sum(state == 0)) 
grp_by_category
```

```{r}
#visualization
ggplot(gofund , aes(y=Category)) + geom_bar(aes(fill=state))
```

```{r}
#Funds
summary(as.numeric(gofund$Raised))
```
```{r}
#Date Difference,state per category
grp_by_category_date <- gofund %>% group_by(Category) %>%
  summarise(date_difference = sum(as.numeric(date_difference)))
grp_by_category_date
```
```{r}
#Visualization
ggplot(gofund, aes(y=Category,x=as.numeric(date_difference))) + geom_col()
```
```{r}
#Word count per category
count_category <- gofund %>% 
  group_by(Category) %>%
  summarise(word_count = sum(as.numeric(WC)))
count_category
```

```{r}
#Word frequency
gofund$Description <- iconv(enc2utf8(gofund$Description),sub="byte")
gofund$text_clean <- tolower(gofund$Description) # Lowercase
gofund$text_clean <- gsub("[[:punct:]]", "", gofund$text_clean) # Remove punctuation

gofund$text_clean <- gsub("[[:digit:]]", "", gofund$text_clean) # Remove numbers
gofund$text_clean <- gsub("\\s+", " ", stri_enc_toutf8(gofund$text_clean)) #Removeextrawhitespaces
vocab <- unlist(str_split(gofund$text_clean, " ")) # Split into vocab list
total_words <- length(vocab) # Total words - 2958646
unique_words <- length(unique(vocab)) # Unique Words - 101764
lex_div <- unique_words / total_words # Lexical Diversity - 0.03439546
vocab_nsw <- vocab[!(vocab) %in% stop_words$word] # Only keep nonstopwords
gofund_text_clean <- paste(vocab_nsw, collapse = " ") # Gather back into one string
freq_word <- sort(table(unlist(strsplit(gofund_text_clean, " "))), #Create frequency table
 decreasing = TRUE)
freq_word <- as.data.frame(freq_word)
freq_word <- freq_word %>%
 rename(word = Var1, freq=Freq)
word_frequency_top10 <- freq_word%>%  top_n(10,freq)

ggplot(word_frequency_top10,aes(word,freq,fill=freq)) +
 geom_col()
```

```{r}

#Word Cloud
word_frequency_top200 <- freq_word%>%
 top_n(200,freq)
wordcloud(words = word_frequency_top200$word, freq =
word_frequency_top200$freq, min.freq = 1,
 max.words=200, random.order=FALSE, rot.per=0.35,
 colors=brewer.pal(8, "Dark2"))
```
```{r}
#Check frequency of words in text
text_df <- subset(gofund, select = c(Description,Category))

text_df %>%
group_by(Category)

```
```{r}
tidy_category <- text_df %>%  unnest_tokens(word, Description)

head(tidy_category)
```

```{r}
#Removing stop words
data(stop_words)

tidy_category <- tidy_category %>%
  anti_join(stop_words)
```
```{r}
#Check frequency of words
tidy_category %>%
  count(word, sort = TRUE) 
```
```{r}
#Visualization
tidy_category %>%
  count(word, sort = TRUE) %>%
  filter(n > 25000) %>%
  mutate(word = reorder(word, n)) %>%
  ggplot(aes(n, word)) +
  geom_col() +
  labs(y = NULL)
```

```{r}
#Effects of posemo and negemo
grp_by_state <- gofund %>% group_by(state) %>%
  summarise(posemo = sum(posemo), negemo = sum(negemo))
grp_by_state

```

```{r}
#nrc emotions related to financial distress,financial strain,financial pressure
nrc <- get_sentiments("nrc")   %>%
  filter(sentiment == c("anger", "fear","sadness","anticip")) 
nrc
```
```{r}
#words linked to financial distress linked,financial strain, financial pressure in the dataset and their nrc per category
tidy_category %>% inner_join(nrc) %>%
  group_by(Category) %>%
  count(word, sort = TRUE) 
```


```{r}
#LIWC output variables related to financial distress, financial stress are negemo,anx,anger,sad,risk,money
#LIWC in the dataset per category

grp_by_category <- gofund %>% group_by(Category) %>%
  summarise(anxiety = sum(anx),anger = sum(anger), sad = sum(sad),risk = sum(risk), money = sum(money), negemo = sum(negemo))
grp_by_category


```