There is More to Sentiment Analysis on Seafood Than Just Crunching Numbers
There are 5 major types of proteins, meat, poultry, seafood, beans and peas; and eggs, processed soy products, nuts, and seeds are considered part of the Protein Foods Group. As dining out has become the norm for celebration/gathering of special occasions and get-togethers, I would say most restaurant-goers would very possibly splurge and pick seafood over meat.
So yes — while the truth is that data drives powerful and meaningful insights, there is so much more to analytics than you would think (especially at how the world is implementing data as a visual aid). Here’s why…As analyzing data is a passion of mine. I am interested in developing a project to find out more about seafood consumer’s behavior and attitude would be.
Sustainable Source of Protein
Hypothesis: Are Seafood(seabass) lovers picky and negatively opinionated?
To give you a “taste” of my thought process, I have put together a series of analysts consisting of graphs that are created using Excel, Python, Matpoltlib, Seaborn, and PostgresSQL. I have applied Yelp’s data and gathered the below analysis.
The challenge that I faced is overcoming data paralysis when there is a massive amount of data at my fingertips. I have searched through 1000 returned results of restaurants with the words “sea bass” in New York City.
Below is my analysis.
My first assumption is usually that the most reviewed restaurants will be closely related to the search terms. However, the most recommended restaurants are displayed on the search results. The above shows the Top 10 cuisines that came up with the words, “sea bass” in New York City. Italian restaurants have the most number of customers’ reviews and followed by American and Japanese in this category
Left: It shows the most number of reviews based in NYC restaurants by zip code that contain the words, “sea bass”.
As the search results by default are sorted through most recommended. I am curious to see if most numbers of reviews have the best results compared to most recommended restaurants. However, it appears that the searches under most recommended returned the best results for the search terms “sea bass”. With words like fish and seafood appearing the most among 300 user reviews. The restaurant “Fish cheek” having the most number of words relating to “Sea bass” & “Fish” appearing.
Interpreting the Numbers & Bringing Them to Life through Impactful Stories
Speaking of data, I was digging further into the user’s sentiment towards a restaurant with the most recommended restaurants. Upon gathering the data, cleaning and all of these tools and data sets are like a million little pieces to a giant puzzle. By identifying the most important pieces of data, synthesizing them, adding context, and bringing them to life through powerful visualizations, and presenting impactful storytelling. I have a list of positive and negative words and matched it against the 300 users’ reviews for the most recommended Restaurant “Fish Cheeks”. It shows that there is an overall positive outcome relating to the following search on Yelp.
There are 257 positive words out of 3091 words. Overall it is showing that customers are mostly happy about the Fish Cheek restaurant.
Findings
In Conclusion, I found out that most restaurants have “delicious” ranks as top 3 among all positive words, indicating that tastes of food matter more than other factors like service or price.
From the negative word list, we could observe that “cold” is one of the main issues for the Fish Cheek restaurant, which we could assume customers expect that type of food to be served warm or hot.
The most reviewed restaurants may not seem to correlate to the search terms whereas the recommended restaurants are.
Based on the sentiment analysis on 1000 restaurants, seafood or seabass lovers actually are not that picky and are spreading more positivity with their choice of words in the reviews overall.