Frequently Asked Questions
Don't See Your Question Here?
Where does Canvs get its data?
Canvs analyzes Facebook, YouTube, Instagram, and Twitter data. Canvs receives its Twitter data from Nielsen, which measures program-related Twitter activity for linear episode airings and on a 24/7 basis. Canvs® is the first platform to automatically and comprehensively measure the emotions that drive conversation and engagement using this data source. You can be confident that the data you see in Canvs is exactly the same data you’re analyzing with Nielsen Twitter TV ratings.
What is an Emotional Reaction [ER]?
Emotional Reactions are defined as any piece of content which contains an emotion. Examples of Emotional Reactions are, “I can’t wait for #PLL,” “That is the scariest zombie ever on Walking Dead,” and “WTF Olivia Pope!” Examples of social media content that do not contain Emotional Reactions are, “I’m watching PLL tonight with my BFF” and “Gotta get back from yoga in time for Scandal.” Canvs displays the volume of Tweets as reported by Nielsen, but Canvs only analyzes Emotional Reactions.
What if a piece of content contains more than one emotional Reaction?
If a piece of content contains more than one Emotional Reaction, Canvs assigns it to the most appropriate cluster. We frequently come across clusters that contain two emotions like Love, Hate because many people have expressed both love and hate for a show in one piece of content. For example, “I love Pretty Little Liars, but I hate Aria.”
EMOTIONAL ANALYSIS METHODOLOGY
How does Canvs create emotional clusters?
Canvs uses our extensive knowledge base to group similar Reactions into the clusters you see in the dashboard. Canvs analytics are based on relationships between words and phrases with similar meanings. We have manually curated the world’s largest knowledge base to understand the way people express emotions on Twitter, so each Emotional Reaction has a relationship to every other Emotional Reaction.
Canvs uses these relationships between Reactions to draw conclusions about which Reactions are similar. For example, someone saying they are "obsessed" with a show means almost the same thing as someone saying they are Addicted to it. Canvs recognizes the close relationship between the ideas Obsessed and Addicted, and is able to group Tweets containing those words with other similar sentiments like “hooked” and “fixated” to produce generalizations about the overall nature of audience Reaction.
What are the possible emotional clusters Canvs produces?
There are hundreds of emotional Reactions that could populate clusters in Canvs. We conducted thorough research and determined that most emotions expressed on Twitter fall into these categories. Some common clusters for TV content include Love, Excited, Hate, Funny, Creepy, Afraid, and Crazy.
How do you keep the Canvs knowledge base up to date?
We have a dedicated team of data scientists who add to our knowledge base every day. They are experts in slang, trends, and vernacular.
How do you handle misspellings?
Misspellings are automatically associated with correct spellings. Every time our analysts identify a misspelling, it gets put into a database that relates it back to the correct spelling of the word. For example, someone commenting the word “coooool” would register to Canvs as “cool.” We’ve seen hundreds of thousands of these misspellings, and we add more every day.
How are the top topics, hashtags, and mentions determined in Canvs?
The topics, hashtags, and mentions are generated directly from Emotional Reactions themselves and ranked in Canvs based on how many times they appear. We are constantly updating Canvs to feature common and show-specific phrases.
What technology powers Canvs?
Canvs leverages a patented, large-scale clustering method that is tailored to handle big data. We have developed multiple breakthrough algorithms under the direction of one of the leading data scientists in America, our Chief Scientist, Professor Sam Hui, PhD, of the University of Houston’s Bauer School of Business.