Word clouds communicate ideas in a simple and engaging way. So why not use them to map sentiment?
People leave their opinions everywhere, in social media conversations, product reviews, surveys, and chats. That’s a lot of information to handle but with sentiment analysis, you can process this data in a matter of minutes.
In this article, we’ll show you how to quickly sort your text data into positive, negative, or neutral using a free sentiment analysis tool, then we’ll show you how to visualize the results using a free word cloud generator.
Word clouds are popular for visualizing qualitative data because they’re simple to use and provide quick insights at a glance.
MonkeyLearn’s free word cloud generator is powered by AI to deliver optimum results. Not only will you see how often words appear, but also how relevant they are. You might also notice phrases (words that often go together) in your word cloud, which help you better understand the context of your data.
Before you create your sentiment word cloud, you’ll first need to parse your text through a sentiment analysis tool.
Follow the steps below to create word clouds with sentiments using free, no-code tools:
First, you’ll need to export your data into a .csv or an Excel file.
For this tutorial, we’ll use product reviews about Slack.
Text data often contains several opinions. In reviews, for example, you’ll often find that a customer is really happy about one feature, but not so happy about another.
To analyze each of these ideas separately (and gain more accurate results), you’ll need to break your dataset into opinion units. Don’t worry, it’s really easy.
First, sign up to use this opinion unit extractor (basically, a tool that extracts and separates opinions). Test the opinion extractor with a paragraph of text to see how it transforms data into more manageable fragments of text, then select “Batch” and upload your Excel or CSV file.
In just a few seconds, you’ll receive a new document with an additional column containing all the extracted opinion units.
The next step is to use MonkeyLearn’s sentiment analyzer to automatically sort each of your opinion units as Positive, Negative, or Neutral.
This is how it works:
To analyze your data in bulk, you’ll need to sign up to MonkeyLearn.
Then, access the model, upload your opinion unit file as a “Batch”, choose the column with the opinion units, and click “Continue”. The model will start analyzing your data and automatically download a new Excel file to your computer, with all your opinions classified by sentiment. Similar to this:
Now that you’ve classified your product reviews by sentiment, it’s time to use those results to create word clouds.
Filter your opinions by sentiment in Excel so you can copy one set of sentiments at a time (positive, neutral, negative). Paste the text into MonkeyLearn’s word cloud generator, just below where it says ‘source text’ and click on “Generate cloud”.
Customize your word cloud by changing color, font, and theme, and choose the number of words you want to appear in your word cloud. You might even want to edit the source text to remove words. Finally, download your word cloud in SVG or PNG.
Here’s a word cloud showing the most frequent words that appear in positive sentiments about Slack:
As you can see, Slack reviews classified as “Positive” frequently contain words like “communications”, “conversation”, “team”, and “channel”. They also mention words like “mobile app”, “user experience”, and “integrations”, specific aspects that are highly valued by Slack users.
But what about negative sentiment?
To find out, repeat the process: filter opinion units by negative sentiment, copy the results, and paste them into the word cloud tool. Here’s a negative sentiment word cloud for Slack:
In this case, Slack reviews with negative sentiment contain common words like “conversation”, “message”, and “communications”. However, less-frequent words or phrases seem to offer more insight, like “service interruption”, “unnecessary email”, “different time zone”, and “paid version”. These keywords give further details about what is frustrating Slack users.
Finally, you can filter product reviews tagged as “Neutral” and create another word cloud to showcase the results, like this:
Neutral opinions are typically more descriptive. Even though they don’t express a particular sentiment, they shed light on topics that are relevant to users.
In this case, keywords like “communication tools”, “file sharing”, “project”, “internal communications”, and “screen sharing”, reveal how customers use the platform.
Learning what customers like and don’t like about your brand can help you get insights that go beyond word frequency, and can even lead to informed decision-making.
Creating word clouds using sentiment analysis data helps visualize how topics are mentioned in your data.
If you want to gain even more granular insights, you can use other text analysis tools, like topic classification.
Sorting your data by topic and sentiment is called aspect-based sentiment analysis and shows you how customers feel about specific aspects of your business.
MonkeyLearn Data Studio is an all-in-one solution that takes you all the way from text analysis to data visualization. Thanks to its intuitive no-code interface, you can easily perform tasks like sentiment analysis, topic analysis, and keyword extraction, and visualize insights through word clouds and charts.
Request a free demo and start using your data to its full potential.
October 1st, 2020