slides_sentiment.Rmd

---
title: "Data for slides"
author: "Dr. Stephen W. Thomas, Queen's University"
date: "2018"
output:
  pdf_document:
    highlight: pygments
    number_sections: yes
    toc: no
    toc_depth: '2'
---



```{r}
library(tidyr)
library(dplyr)
library(readr)
```

# Data

```{r}
sentences = c(
  'I am happy.',
  'I am very happy.',
  'I am not very happy.',
  'I love cheese but I hate wine.',
  'Nobody gives a good performance.',
  'She is sleazy.',
  'She is slightly sleazy.',
  'GOING TO LA THIS WEEKEND.',
  'The boy is not mad.',
  'The boy is not not mad.',
  'Princess Elsa is a beautiful person to me.',
  'Princess Elsa is not a beautiful person to me.',
  'Princess Elsa is a very beautiful person to me.',
  'Princess Elsa is not a very beautiful person to me.',
  'I love being ignored.'
)
```

The results data frame.

```{r}
res = data.frame(sentence = character(), meth = character(), sentiment=numeric())
```

# QDAP

```{r}
library(qdap)

res1 = data.frame(sentence = character(), meth = character(), sentiment=numeric())
for (s in sentences){
  p = polarity(s, n.before=4, n.after=2)
  res1 = rbind(res1, data.frame(sentence=s, meth="qdap", sentiment=p$all$polarity))
  #p = polarity(s, n.before=2, n.after=2)
  #res = rbind(res, data.frame(sentence=s, meth="qdap 2/2", sentiment=p$all$polarity))
  #p = polarity(s, n.before=2, n.after=4)
  #res = rbind(res, data.frame(sentence=s, meth="qdap 2/4", sentiment=p$all$polarity))
  #p = polarity(s, n.before=4, n.after=4)
  #res = rbind(res, data.frame(sentence=s, meth="qdap 4/4", sentiment=p$all$polarity))
  #p = polarity(s, n.before=6, n.after=4)
  #res = rbind(res, data.frame(sentence=s, meth="qdap 6/4", sentiment=p$all$polarity))
}
res1

#qdapDictionaries::amplification.words
#polarity("cheese great")
#polarity("barely great")

res = rbind(res, res1)
res1

```

# Syuzhet


```{r}
library(syuzhet)


res1 = data.frame(sentence = character(), meth = character(), sentiment=numeric())
for (s in sentences){
  res1 = rbind(res1, data.frame(sentence=s, meth="syuzhet", sentiment=syuzhet::get_sentiment(s)))
  #res1 = rbind(res1, data.frame(sentence=s, meth="syuzhet, bing", sentiment=syuzhet::get_sentiment(s, method="bing")))
  #res1 = rbind(res1, data.frame(sentence=s, meth="syuzhet, afinn", sentiment=syuzhet::get_sentiment(s, method="afinn")))
  #res1 = rbind(res1, data.frame(sentence=s, meth="syuzhet, nrc", sentiment=syuzhet::get_sentiment(s, method="nrc")))
}

res = rbind(res, res1)
res1
```

# SentimentAnalysis

```{r}
library(SentimentAnalysis)


res1 = data.frame(sentence = character(), meth = character(), sentiment=numeric())
for (s in sentences){
  res1 = rbind(res1, data.frame(sentence=s, meth="SentimentAnalysis", sentiment=SentimentAnalysis::analyzeSentiment(s)$SentimentGI))
}


res = rbind(res, res1)
res1
```

# RSentiment

```{r}
library(RSentiment)


res1 = data.frame(sentence = character(), meth = character(), sentiment=numeric())
for (s in sentences){
  res1 = rbind(res1, data.frame(sentence=s, meth="RSentiment", sentiment=RSentiment::calculate_sentiment(s)$sentiment))
}


res = rbind(res, res1)
res1
```

# Tidy

```{r}
library(tidytext)

df = data_frame(text=sentences) %>%
  mutate(docid = row_number())


tidy <- df %>% 
  unnest_tokens(word, text)

res1 = tidy %>%
  inner_join(get_sentiments("afinn"), by=c("word")) %>%
  group_by(docid) %>%
  summarise(sentiment=sum(score)) %>%
  inner_join(df, by=c("docid")) %>%
  mutate(sentence=text) %>%
  mutate(meth="tidy") %>%
  select(sentence, meth, sentiment)

res = rbind(res, res1)
res1
```


# sentimentr

```{r}
library(sentimentr)


res1 = data.frame(sentence = character(), meth = character(), sentiment=numeric())
for (s in sentences){
  res1 = rbind(res1, data.frame(sentence=s, meth="sentimentr", sentiment=sentimentr::sentiment(s)$sentiment))
}


res = rbind(res, res1)
res1
```



# Azure

Gathered Sept 11, 2018: https://azure.microsoft.com/en-us/services/cognitive-services/text-analytics/

```{r}
res1 = data.frame(sentence = character(), meth = character(), sentiment=numeric())

res1 = res1 %>%
  add_row(sentence='I am happy.', meth="Azure", sentiment=1) %>%
  add_row(sentence='I am very happy.', meth="Azure", sentiment=1) %>%
  add_row(sentence='I am not very happy.', meth="Azure", sentiment=.01) %>%
  add_row(sentence='I love cheese but I hate wine.', meth="Azure", sentiment=.15) %>%
  add_row(sentence='Nobody gives a good performance.', meth="Azure", sentiment=.87) %>%
  add_row(sentence='She is sleazy.', meth="Azure", sentiment=.76) %>%
  add_row(sentence='She is slightly sleazy.', meth="Azure", sentiment=.76) %>%
  add_row(sentence='GOING TO LA THIS WEEKEND.', meth="Azure", sentiment=.85) %>%
  add_row(sentence='The boy is not mad.', meth="Azure", sentiment=.99) %>%
  add_row(sentence='The boy is not not mad.', meth="Azure", sentiment=.26) %>%
  add_row(sentence='Princess Elsa is a beautiful person to me.', meth="Azure", sentiment=.92) %>%
  add_row(sentence='Princess Elsa is not a beautiful person to me.', meth="Azure", sentiment=.14) %>%
  add_row(sentence='Princess Elsa is a very beautiful person to me.', meth="Azure", sentiment=.92) %>%
  add_row(sentence='Princess Elsa is not a very beautiful person to me.', meth="Azure", sentiment=.14) %>%
  add_row(sentence='I love being ignored.', meth="Azure", sentiment=.81)
  
res = rbind(res, res1)
res1
```


# Comparison

```{r}
comparison = res %>%
  spread(meth, sentiment)
comparison
```


# Emotions


```{r}
library(syuzhet)

s ='I am actually very happy today.'
emo=syuzhet::get_nrc_sentiment(sentences)
emo = cbind(sentences, emo)
emo
```