Facebook reactions analysis

Facebook is without any doubt the most popular social media platform in the world. We comment, like and share stuff every day. We interact with others in different ways, be it from sending messages, commenting or putting a “Like” on a post or a comment.

In 2016, Facebook redesigned the “Like” button and added the “Love,” “Haha,” “Wow,” “Sad,” and “Angry” buttons. It opened a window to the users feelings, that was a pretty smart move from Facebook especially in this day and age with data analytics gaining more ground.

As we know, human behavior is hard to contain. It is highly variable and relative from person to person. No one equation can define human interactions, not even specific ones like joy or anger. Someone could be happy for some reasons whilst another would be sad or angry for that exact same reason, that shows how complex human behavior is.

Here, we’ll try to get a glimpse of how users react to French and English news on Facebook from the data we have from all the “Loves“, “Hahas“, “Wows“, “Sads” and “Angrys” (excluding the “Like” because it doesn’t reflect any specific emotional state). We’ll be doing an exploratory analysis, which basically means that we don’t know what to expect from the data. We could end with something that’s relevant as we could end up with something that’s irrelevant, especially with data from social medias. We will also not take into consideration specific news subjects or biases.

On Facebook, users can either drop a ‘Like‘ (‘reactions’ in our case), a comment or share a post. Which is why we’ll see how ‘reactions’ relate to those 3 ‘actions’.

Which are:

  • Reactions: the sum of all ‘Loves‘, ‘Wows‘, ‘Hahas‘, ‘Sads‘ and ‘Angrys
  • Comments: the amount of comments
  • Shares: the amount of shares

Our ‘reactions’ are, ‘”Loves“, “Hahas“, “Wows“, “Sads” and “Angrys“.

We got two main groups, French Canadians and English Canadians. There’s also the bilinguals that know both languages. In Canada, approximately less than one fifth speak both French and English. (We won’t be going into that …)

Let’s say we wanted to get more ‘Likes’, ‘Shares’ or ‘Comments’, (i.e traffic) on our Facebook page and we aimed at a specific group of users, what would you do ? The first logical step to take is to see how those same users act on Facebook, right ? Which is why in our case we’ll be scraping media outlets pages, they are filled with daily news posts, from bad to good to get a full spectrum of all users ‘reactions’.

Which brings us to our question:

  • What are the sentiments that motivates French and English users to react, comment and share a post ?


After a thorough cleaning and some wrangling, we end up with that plot.


Screenshot_2018-05-21 Facebook sentiment analysis
French sentiment correlation
Screenshot_2018-05-21 Facebook sentiment analysis(1)
English sentiment correlation


So the simple answer is:

French users tend to react a lot with ‘Sad’ posts and English users with ‘Love’ posts.

The main sentiment that tend to motivate French users to comment is ‘Sad’, and for English users it is ‘Angry’.

The main reaction that motivates French users to share a post is ‘Sad’, and fore English users it is ‘Love’.

But it isn’t as easy as it seems, as i said in the beginning human behavior is really hard to encapsulate and understanding it fully is nearly impossible due to multiple internal and external factors. But with the new age of data we can get a glimpse of it.


Full project code —> here

Credit: Scraping was done with code available on Github, thanks to minimaxir !


Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s