New in Awario: Sentiment Analysis for every language
Awario has offered sentiment analysis for a few years now, but it only worked for English and - let's face it - it wasn't perfect. That changes today - meet the brand new sentiment algorithm that works for every language and is available in every pricing plan.
What is sentiment analysis?
Sentiment analysis is the process of determining the attitude expressed by someone towards a topic or phenomenon. In the world of social listening, sentiment analysis is most commonly conducted for online mentions of a brand, and is used to quantify the attitude of consumers towards that brand as part of brand health analysis. Most sentiment analysis tools categorise mentions into positive, negative, and neutral.
Awario not only identifies the sentiment of your mentions, but it also comes up with hypotheses (aka Insights) for reasons behind any spikes in the volume of positive or negative conversations. As you can see in the image above, a spike in negative mentions for Airbus has been caused by mentions containing words like tariff and tariffs; by clicking on these insights you can dig deeper into the mentions and see what caused the influx of negative conversations.
Sentiment analysis is also an important part of market research and competitor analysis. By looking at consumer sentiment towards your own brand you won't always be able to say if you're doing good or bad; it's benchmarking those values against the competition that lets you put everything in perspective.
What's new in Awario's sentiment analysis?
We've completely rebuilt Awario's sentiment algorithm to bring you data that's more accurate and comprehensive. So what exactly is new in today's update?
1. Support for every language
You read that right - the new sentiment algorithm works for each and every language out there! We've done extensive testing for the languages that are most popular with Awario users (English, French, Spanish, German, and Portuguese) as well as smaller tests for other languages, and are happy to say that the accuracy of sentiment analysis for no language falls below 65%.
2. A new, better algo
We've tested several new approaches to analyzing sentiment with thousands of human evaluators to determine the accuracy of each one. We came up with a system that scored over 70% across all languages. Bear in mind that the accuracy of sentiment analysis can never reach 100% - even for humans, the agreement rate on sentiment of social media posts has been shown to be around 65%.
Here are just a few examples to give you an idea of how tricky determining sentiment can be for machines.
With little context, it can be very hard to identify if a document is sarcastic even for humans. Here's one less subtle example where the sentiment might be obvious for humans, but not algorithms.
It only took me 5 minutes to get a coffee at Starbucks. Great start of the day!
It only took me 30 minutes to get a coffee at Starbucks. Great start of the day!
Emotional words, such as love and hate, are easy to interpret to both humans and machines. However, some words can be negative in one context, and neutral or positive in another, such as in the example below.
I tend to drink ice cold coffee in the summer.
When I finally got my coffee, it was ice cold.
As I'm sure you can see, sentiment analysis is a complicated process. We have an in-depth guide coming up on the use cases of sentiment analysis and the specifics on how it works. Make sure you're subscribed to our newsletter so that you don't miss it!
Excited to take the improved sentiment algo for a test drive? Go to your Awario dashboard and share your thoughts in the comments below!