5 Real World Sentiment Analysis Examples

Whether positive or negative, opinions are powerful. In the digital world, they can boost or ruin a brand’s reputation. Stats show that 40% of buyers form an opinion of a business after reading 1-3 online reviews, giving us a clue of how important it is for companies to track the conversation around them and uncover the feelings behind what’s being said.

Dealing with the huge amounts of data businesses generate daily can be a struggle when done manually. That's when sentiment analysis comes in, this technology allows you to automatically identify the emotional tone in a text. Thanks to Natural Language Processing (NLP), it is possible to create systems that are able to understand the opinions present in all kinds of conversations and obtain valuable insights about products or services.

In this guide, we’ll dive into the different applications of sentiment analysis and provide use cases and examples to help you understand the concept. We'll also show how easy it is to create a sentiment analysis model with a SaaS tool like MonkeyLearn.

  1. What is sentiment analysis
  2. Examples of sentiment analysis
  3. Types of sentiment analysis
  4. Creating a custom sentiment analysis classifier

Let’s get started!

What is Sentiment Analysis

Sentiment analysis is an automated process capable of understanding the feelings or opinions that underlie a text. It is one of the most interesting subfields of NLP, a branch of Artificial Intelligence (AI) that focuses on how machines process human language.

Sentiment analysis studies the subjective information in an expression, that is, the opinions, appraisals, emotions, or attitudes towards a topic, person or entity. Expressions can be classified as positivenegative, or neutral. For example:

  • “I really like the new design of your website!” → Positive
  • “I’m not sure if I like the new design” → Neutral
  • “The new design is awful!” → Negative

Sentiment analysis is a powerful tool that can have a great impact in many business areas, like social media management, marketing, product, and customer support. Keep reading to learn about the most interesting sentiment analysis business applications.

Examples of Sentiment Analysis

Social Media Monitoring

We live in a world where 3.5 billion people are active social media users; that’s 45% of the world’s population! Every single minute of the day, people write more than 500,000 Tweets and 510,000 Facebook comments. This immense amount of data is a gold mine of invaluable information on people’s opinions, preferences, and emotions towards a myriad of things: from the mobile phone that they wish to buy to the political candidate that they intend to vote for.

Sentiment analysis allow us to mine this data and extract the feelings that underlie social media conversation, to understand how people are talking about a given product or topic.

An example of a negative tweet

The tweet above is an example of a negative comment. According to stats, 46% of people have used social media to escalate their complaints towards a company. Thanks to sentiment analysis, we are able to automatically identify those dissatisfied customers, categorize their issues by urgency, and prioritize their requests.

Real-time sentiment analysis can put you one step ahead of any potential PR crisis, allowing you to take action before a customer’s bad experience goes viral.

Finally, you can use sentiment analysis to analyze your competition. You can track how customers are talking about them and find opportunities to improve your own business.

This sentiment analysis model for tweets about airline comments is an example of sentiment analysis for social media monitoring. You can paste a tweet mentioning @AmericanAir or @VirginAtlantic in the box and it will be automatically tagged as positive, negative, or neutral:

Test with your own text



Brand Monitoring

Besides social media, online conversations can take place in blogs, review websites, news websites and forum discussions. Product reviews, for instance, have become a crucial step in the buyer decision process. Consumers read at least ten reviews before buying, and 57% will only trust a business if it’s rated with 4 or more stars.

Sentiment analysis is an excellent tool to keep a close eye on your brand’s reputation, find out what is right or wrong about your business, and understand more about your customers.

If you need to get detailed insights on different features related to your product, you should try aspect-based sentiment analysis. This will allow you to see what specific things about your product are being praised or criticized by your customers.

For example, let’s take a few Drift reviews and see how an aspect-based model would classify them:

Drift reviews

Sentiment analysis can also be particularly helpful to monitor online conversation on a specific point in time, for example, if you are launching a new product, releasing a new update or starting a new marketing campaign.

Ready to see how it works by yourself? Here’s a pre-trained sentiment analysis model for product reviews:

Test with your own text



Customer Support

Providing outstanding customer service experiences should be a #1 priority for a successful company. After all, 96% of consumers say great customer service is a key factor to choose and stay loyal to a brand.

Fortunately, sentiment analysis can help you make your customer support interactions faster and more effective.

If you run sentiment analysis on all your incoming tickets, you would be able to easily detect the most dissatisfied customers or the most urgent issues and prioritize them above the rest. Plus, you could route tickets to the appropriate person or team in charge of dealing with them.

You can also use sentiment analysis to assess the results of your customer support strategy. Let’s take this Tweet complaining about Airbnb customer support:

Tweet complaining about Airbnb

By analyzing sentiment on customer support chats or tweets referring to interactions with your customer support team (like the one above), you can get relevant insights regarding customer satisfaction and detect clear opportunities for improvement.

Customer Feedback

Net Promoter Score (NPS) surveys are one of the most common ways of knowing how customers perceive a product or service. Basically, they consist of two stages: first, you ask a customer to score a business from 0 to 10, then you ask them to give reasons for the score they leave with open-ended question.

When it comes to processing the results, the first stage is easy: you just have to calculate the average score. But when it comes to analyzing tons of open-ended NPS responses, the analysis becomes more complicated. Imagine if your team had to tag hundreds of responses manually. Not only it would be a tedious and time-consuming task, it may also lead to inconsistent results derived from different criteria during the tagging process.

Fortunately, sentiment analysis enables you to process large volumes of NPS responses and obtain consistent results in a very fast and simple way.

The scores in NPS surveys allow you to classify customers as promoters, passives, and detractors. Thanks to sentiment analysis, you can go beyond the numbers to identify the reasons for NPS scores.

Let’s say a customer gave your business a score of ‘7’, and then added: 'The product is decent, but your website is so confusing it took me forever to find the product I was looking for'. In this case, the second part of the survey is the one that contains the most valuable information, pointing out that you should consider improving your website’s UX-UI.

By running an aspect-based sentiment analysis on a set of open-ended NPS responses, you’ll gauge sentiments regarding specific features of your product. That way, you’ll find out what customers appreciate and dislike most about your product. Once you’ve got a sentiment analysis process up and running, you’ll also be able to compare results with previous NPS surveys and see how sentiment surrounding aspects of your product has improved over time.

Market Research

One last application of sentiment analysis has to do with market research.

If you are planning to release a new product, and you want to collect insights on people’s attitudes, experiences, and needs related to the field of your product, you could use sentiment analysis to track different sources of online conversations about a given topic. The same thing applies for following new trends.

Let’s take the case of foldable phones, for example:

Tweet about foldable phones

Another tweet about foldable phones

Performing sentiment analysis of tweets referring to the topic ‘foldable phones’ will allow you to understand how people feel about this product: do they find it exciting? Do they think it may be a useful product? Would they buy it?

Types of Sentiment Analysis

As you can see from the previous examples, it is possible to build sentiment analysis models oriented to different purposes. Even though the most common type of sentiment analysis focuses on polarity (classifying an opinion as positive, negative, or neutral), other types may focus on detecting feelings, emotions, or intentions.

These are the most common types of sentiment analysis:

Standard Sentiment Analysis

It identifies the nuance of an opinion and classifies it as PositiveNegative, or Neutral. It’s the most popular type of sentiment analysis. For example:

  • 'I love how Zapier takes different apps and ties them together' → Positive
  • “I still need to further test Zapier to say if its useful for me or not' → Neutral
  • “Zapier is sooooo confusing to me' → Negative

Fine-grained Sentiment Analysis

Also focused on polarity, this type of sentiment analysis adds a few more categories to obtain more granular results. Similar to 5-star ratings, it classifies opinions as:

  • Very positive
  • Positive
  • Neutral
  • Negative
  • Very negative

For example, imagine having the following survey responses:

  • 'The older interface was much simpler' → Negative
  • 'Awful experience. I would never buy this product again!' → Very Negative
  • 'I don't think there is anything I really dislike about the product' → Neutral

Emotion Detection

This sentiment analysis model detects the emotions that underlie a text. It makes associations between words and emotions like anger, happiness, frustration, etc. For example,

  • 'Hubspot makes my day a lot easier :)' → Happiness
  • 'Your customer service is a nightmare! Totally useless!!' → Anger

Aspect-based Sentiment Analysis

This type of sentiment analysis focuses on understanding the aspects or features that are being discussed in a given opinion. Product reviews, for example, are often composed of different opinions about different characteristics of a product, like Price,UX-UIIntegrationsMobile Version, etc. Let’s see some examples:

HubSpot Example

SurveyMonkey Example

Intent Detection

This type of sentiment analysis tries to find an action behind a given opinion, something that the user wants to do. Identifying user intents allows you to detect valuable opportunities to help customers, such as solving an issue, making improvements on a product or deriving complaints to the correspondent areas:

  • “Very frustrated right now. Instagram keeps closing when I log in. Can you help?” → Request for Assistance

Customers experiencing issues can be easily spotted thanks to sentiment analysis.

Creating a Custom Sentiment Analysis Classifier

We’ve introduced you to a few examples of pre-trained models for sentiment analysis. These models are a great option if you want to get started right away. Plus, you can upload your own data into the models and obtain results in a very simple and fast way. If you want to see all the models available, just register on MonkeyLearn and click on ‘Explore’.

However, if you need more accuracy, we recommend you create a custom sentiment analysis classifier. That way, you can use specific data from your field to train it and make sure it follows your own criteria. MonkeyLearn allows you to create custom models with machine learning.

To get started with the process, just follow these steps:

1.  Go to ‘create a model’ and choose ‘classifier’:

MonkeyLearn's creation wizard, with the option to choose a classifier or extractor.

2. Then, choose ‘sentiment analysis’:

MonkeyLearn's creation wizard showing classification options: topic, sentiment, and intent.

3. Import your text data

You can import data from several sources, like Excel, CSV, Twitter, Gmail, and Zendesk, among others:

MonkeyLearn's creation wizard showing available sources to upload your text data.

4. Start training your model by tagging each sentence with the appropriate category.

Tag data to train your model

5. Test your model

Once you have tagged a certain amount of samples, the platform will ask you to name your new model and give you the option to test it or keep training it:

New sentiment analysis model

Once you’ve named your model, you start testing it to see how much your model has learned from the examples you provided, or keep training it to improve accuracy:

Test your model

6. Put the model to work!

Once you’ve finished training your sentiment analysis classifier, it’s ready to analyze new data. The next step is to process the data you want to analyze with the model you’ve just created. With MonkeyLearn, there are three ways you can add new data :

  • Upload an Excel or CSV file and classify data in a batch.
  • Use one of the integrations to connect MonkeyLearn with a third-party app (Zapier, Google Sheets, Rapid Miner, among others).
  • Use the API with your favorite programming language.

Final Words

Sentiment analysis provides amazing insights on customers’ feelings and opinions.

What are the things that people like or dislike about your brand? What are their favorite features of your product? Which aspects require more improvement? How do they feel about your company’s customer service? These are just a few of the many questions that can be answered through sentiment analysis.

The possibilities of machines capable of understanding the human language are endless, and thanks to platforms like MonkeyLearn, you don’t need to be a machine learning expert to get a sentiment analysis model up and running.

Inés Roldós

July 20th, 2020