Skyline vs. Gold Star Sentiment Analysis

Author

Samantha Farfsing

Yelp - Skyline vs. Gold Star Sentiment Analysis

As someone that has lived in Cincinnati my entire life, I know that the debate over what chili restaurant truly represents the “Cincinnati flavor”. In my opinion, Skyline is the superior choice but I decided to use relevant Yelp reviews in order to prove how customers feel about the two options. With this data, I will perform a sentiment analysis and answer the following questions about the data:

Are customers of Skyline generally more satisfied with their experience compared to Gold Star?
How have the sentiment scores for each restaurant varied throughout time?
How does the emotive sentiment scoring differ for each restaurant?

Obtaining and Loading the Data

The combined data from scraping both of these Yelp Review pages can be found with this link:

https://myxavier-my.sharepoint.com/:x:/g/personal/farfsings2_xavier_edu/EeIs4lIZIsJNsbTNRkHzSb8B2CpEsGtziCa1pJy4rAQrlA?download=1

The packages required for this analysis are:

# Loading Packages
library(tidyverse)
library(rvest)
library(lubridate)
library(httr)
library(tidytext)

Data Dictionary

Reviewer Name
Reviewer Location
Review Date
Review Rating
Review Content

Analysis

Are customers of Skyline generally more satisfied with their experience compared to Gold Star?

To answer this question, I collected data from the Yelp reviews of each restaurant in order to find the count of positive and negative words for each. After collecting this data, I used bing to review the counts of positive and negative words for each restaurant. Then, I created a visual to show the comparisons between the two. This data will show the frequency of positive and negative words found in the Yelp reviews.

The graph shows that both restaurants seem to have higher frequency of words that would be considered ‘positive’, compare to ‘negative’ ones. Overall, Skyline has more words that appear, especially on the ‘positive’ side. This could help indicate how customers feel, but further analyis is definitely needed before determining what customers prefer.

How have the sentiment scores for each restaurant varied throughout time?

To answer this question, I used the data to find the average review score for each restaurant from 2005-now. This analysis will help show how customer reviews have changed throughout the years for the two restaurants, and if a certain restaurant holds higher average reviews overall per year than the other.

Based on this line chart, both restaurants have seen significant falls in average review scores at various times over the years. The restaurants follow a fairly close trend until around 2014 when Gold Star’s average rating went down to 2 while Skyline’s stood at 3.5. Following this, Skyline’s average stayed fairly steady, while Gold Star’s saw both increases and decreases. The graph shows that Gold Star’s average took another fall last year, in 2023, but is back on the rise. On the contrary, Skyline seems to be seeing a decrease in their reviews currently. In my opinion, this is most likely due to the recent change from Pepsi products to Coca-Cola. I would be interested to re-collect the data after all locations have implemented this change to see how these reviews change, and if Gold Star’s reviews increase as a result.

How does the emotive sentiment scoring differ for each restaurant?

To answer this question, I used the collected data along with the NRC lexicon to score the various emotive words found within each restaurants’ reviews. I used these scores to create a bar graph to compare both restaurants and the count of words that fit into each category of the NRC lexicon. This will help to show the different emotions that customers feel in their reviews for each restaurant.

After looking at the visual, it is very clear that both restaurants show a higher total number of positive words compared to negative ones. The total for positive words within both restaurants’ reviews are very close. For negative words, however, Skyline seems to have a higher total number compared to Gold Star. Also, under anger, disgust, and sadness, Skyline also has higher total numbers than Gold Star. However, the totals under joy are fairly close in total.

Conclusion

After collecting the Yelp review data and performing various analyses on it, it seems as though Gold Star overall has higher sentiment ratings compared to Skyline. The average review scores over time vary for both restaurants, so it is difficult to determine if this is an indication of which chili restaurant truly holds the heat of Cincinnatians. There is much more analysis to be done on this topic, and various locations for each restaurant that could impact the scores. From this data, Gold Star seems to take the win, but it would be interesting to add more elements into the analysis. Although this is the result of my analysis, Skyline will always be in winner in my eyes.

--- title: "Skyline vs. Gold Star Sentiment Analysis" author: "Samantha Farfsing" editor: visual format: html: code-tools: TRUE # Allow the code tools option showing in the output. self-contained: TRUE # Embeds all components into a single HTML file. execute: warning: FALSE # Prevent all warnings from displaying in the output. message: FALSE # Prevent all messages from displaying in the output. --- ## Yelp - Skyline vs. Gold Star Sentiment Analysis As someone that has lived in Cincinnati my entire life, I know that the debate over what chili restaurant truly represents the "Cincinnati flavor". In my opinion, Skyline is the superior choice but I decided to use relevant Yelp reviews in order to prove how customers feel about the two options. With this data, I will perform a sentiment analysis and answer the following questions about the data: 1. Are customers of Skyline generally more satisfied with their experience compared to Gold Star? 2. How have the sentiment scores for each restaurant varied throughout time? 3. How does the emotive sentiment scoring differ for each restaurant? ## Obtaining and Loading the Data The combined data from scraping both of these Yelp Review pages can be found with this link: <https://myxavier-my.sharepoint.com/:x:/g/personal/farfsings2_xavier_edu/EeIs4lIZIsJNsbTNRkHzSb8B2CpEsGtziCa1pJy4rAQrlA?download=1> The packages required for this analysis are: ```{r, echo=TRUE, results='hide'} # Loading Packages library(tidyverse) library(rvest) library(lubridate) library(httr) library(tidytext) ``` ```{r, echo=FALSE, include=FALSE} combined_reviews <- read_csv("combined_reviews.csv") ``` ## Data Dictionary - Reviewer Name - Reviewer Location - Review Date - Review Rating - Review Content ## Analysis ### Are customers of Skyline generally more satisfied with their experience compared to Gold Star? To answer this question, I collected data from the Yelp reviews of each restaurant in order to find the count of positive and negative words for each. After collecting this data, I used bing to review the counts of positive and negative words for each restaurant. Then, I created a visual to show the comparisons between the two. This data will show the frequency of positive and negative words found in the Yelp reviews. ```{r, echo=FALSE, include=FALSE} # Create a review ID variable based on the date combined_reviews <- combined_reviews %>% arrange(review_date) %>% mutate(review_id = row_number()) # unnest tokes by location and remove stop words tidy_reviews <- combined_reviews %>% unnest_tokens(word, review_content) %>% anti_join(stop_words) # simple word counts for each location tidy_reviews %>% group_by(restaurant_name, word) %>% summarize(n = n()) %>% arrange(-n) bing <- get_sentiments("bing") combined_counts <- tidy_reviews %>% group_by(restaurant_name, word) %>% summarize(n = n()) %>% inner_join(bing) combined_counts %>% group_by(restaurant_name,sentiment) %>% summarize(n = n()) %>% arrange(-n) ``` ```{r, echo=FALSE} combined_counts %>% filter(!word == 'gold') %>% filter(n>5) %>% mutate(n = ifelse(sentiment == "negative", -n, n)) %>% mutate(word = reorder(word, n)) %>% ggplot(aes(word, n)) + geom_col() + coord_flip() + facet_wrap(~restaurant_name, ncol = 2) + geom_text(aes(label = signif(n, digits = 3)), nudge_y = 8) + labs(title = "Positive and Negative Words for Gold Star and Skyline", subtitle = "Only words appearing at least 5 times are shown", x = "Count", y = "Word") ``` The graph shows that both restaurants seem to have higher frequency of words that would be considered 'positive', compare to 'negative' ones. Overall, Skyline has more words that appear, especially on the 'positive' side. This could help indicate how customers feel, but further analyis is definitely needed before determining what customers prefer. ### How have the sentiment scores for each restaurant varied throughout time? To answer this question, I used the data to find the average review score for each restaurant from 2005-now. This analysis will help show how customer reviews have changed throughout the years for the two restaurants, and if a certain restaurant holds higher average reviews overall per year than the other. ```{r, echo=FALSE} combined_reviews <- combined_reviews %>% mutate(year = year(review_date)) average_scores <- combined_reviews %>% group_by(year, restaurant_name) %>% summarize(average_review_score = mean(review_rating)) ggplot(average_scores, aes(x = year, y = average_review_score, color = restaurant_name)) + geom_line() + labs(title = "Average Review Scores Over Time", x = "Year", y = "Average Review Score", color = "Restaurant Name") ``` Based on this line chart, both restaurants have seen significant falls in average review scores at various times over the years. The restaurants follow a fairly close trend until around 2014 when Gold Star's average rating went down to 2 while Skyline's stood at 3.5. Following this, Skyline's average stayed fairly steady, while Gold Star's saw both increases and decreases. The graph shows that Gold Star's average took another fall last year, in 2023, but is back on the rise. On the contrary, Skyline seems to be seeing a decrease in their reviews currently. In my opinion, this is most likely due to the recent change from Pepsi products to Coca-Cola. I would be interested to re-collect the data after all locations have implemented this change to see how these reviews change, and if Gold Star's reviews increase as a result. ### How does the emotive sentiment scoring differ for each restaurant? To answer this question, I used the collected data along with the NRC lexicon to score the various emotive words found within each restaurants' reviews. I used these scores to create a bar graph to compare both restaurants and the count of words that fit into each category of the NRC lexicon. This will help to show the different emotions that customers feel in their reviews for each restaurant. ```{r, echo = FALSE} nrc <- get_sentiments("nrc") tidy_reviews %>% inner_join(nrc, by = "word", relationship = "many-to-many") %>% group_by(sentiment, restaurant_name) %>% summarize(n = n()) %>% ggplot(aes(x = sentiment, y = n, fill = restaurant_name)) + geom_bar(stat = "identity", position = "dodge") + scale_fill_manual(values = c("red","#002393")) + labs(title = "Restaurant Sentiment Scores", subtitle = "Total number of emotive words scored ", y = "Total Number of Words", x = "Emotional Sentiment", fill = "Restaurant") # Create a data frame of sentiment scores for each review sentiments <- tidy_reviews %>% inner_join(nrc, by = "word", relationship = "many-to-many") %>% group_by(sentiment, review_id) %>% summarize(n = n()) %>% pivot_wider(names_from = sentiment, values_from = n, values_fill = 0) %>% mutate(positivity = positive-negative) restaurants_sentiments <- left_join(combined_reviews, sentiments, by = "review_id") %>% replace(is.na(.),0) ``` After looking at the visual, it is very clear that both restaurants show a higher total number of positive words compared to negative ones. The total for positive words within both restaurants' reviews are very close. For negative words, however, Skyline seems to have a higher total number compared to Gold Star. Also, under anger, disgust, and sadness, Skyline also has higher total numbers than Gold Star. However, the totals under joy are fairly close in total. ## Conclusion After collecting the Yelp review data and performing various analyses on it, it seems as though Gold Star overall has higher sentiment ratings compared to Skyline. The average review scores over time vary for both restaurants, so it is difficult to determine if this is an indication of which chili restaurant truly holds the heat of Cincinnatians. There is much more analysis to be done on this topic, and various locations for each restaurant that could impact the scores. From this data, Gold Star seems to take the win, but it would be interesting to add more elements into the analysis. Although this is the result of my analysis, Skyline will always be in winner in my eyes.