Understanding Facebook Reactions using Google Sentiment Analysis

Florin Badita
Dec 24, 2016 · 8 min read

2019update: New article out of the owen:

Image for post
Image for post

If you don`t know what Sentiment Analysis is, you can read a article written by Matt Kiser here:

I decided to run Google Cloud Sentiment Analysis over the facebook posts that i`m monitoring using , a amazing tool that i made for journalists that want to eliminate fake viral news.

I wanted to see if we can get a better understanding of the posts that are polarized with a specific Facebook reaction (love,haha,sad,angry,wow).

Image for post
Image for post

For example, for , i have 261 total reactions.

And 129 Likes. I subtract the number of Likes from Total Reactions and i remain with the total number of “emotional reactions”.

Them, i compute for each emotion, the percentage.

Image for post
Image for post

Part 2

Processing over 20K of Facebook statuses using G Cloud Natural Language API and writing the result in a separate model created to hold the data.

To be able to understand the results, we first need to understand what score and magnitude means:

Interpreting sentiment analysis values

The score of a document’s sentiment indicates the overall emotion of a document. The magnitude of a document’s sentiment indicates how much emotional content is present within the document, and this value is often proportional to the length of the document.

A document with a neutral score (around 0.0) may indicate a low-emotion document, or may indicate mixed emotions, with both high positive and negative values which cancel each out.
Generally, you can use magnitude values to disambiguate these cases, as truly neutral documents will have a low magnitude value, while mixed documents will have higher magnitude values.

  • score of the sentiment ranges between -1.0 (negative) and 1.0 (positive) and corresponds to the overall emotional leaning of the text.
  • magnitude indicates the overall strength of emotion (both positive and negative) within the given text, between 0.0 and +inf. Unlike score, magnitude is not normalized; each expression of emotion within the text (both positive and negative) contributes to the text's magnitude (so longer text blocks may have greater magnitudes).
  • The chart below shows some sample values and how to interpret them:
Image for post
Image for post

Part 3. Profit

I find it fascinating to be able to see Facebook Emotions being clustered around the score and magnitude axes that Google Sentiment Analysis predicted for each of the FB posts text.

Image for post
Image for post

Facebook Angry Reactions — Sentiment Analysis

Image for post
Image for post

We can see for the posts that generated a strong Angry Response, the majority of them are classified by google cloud as texts with a negative or neutral score.

There are angry posts that have a positive score, but the majority of posts that generated a angry response, have a text description that got a negative score.

If we look also at the magnitude, we see that it`s getting harder to find angry posts in the positive score part (right part of the image), after the magnitude of the text gets bigger then 1.5.

Facebook Love Reactions — Sentiment Analysis

Love is positive,hate is negative. Proven with data.

Image for post
Image for post

If we look were is the love, we see a different distribution. Articles that have a negative score tend to get less love. At least if the magnitude of the text is over 1

Wows posts are shallow and superficial

I never understood the wows. What are they ? It`s not a haha, is not a like, is not love. I consider them something shallow . I don`t get wows.
Why i`m saying this story ?

Because this play good with my that wow is superficial,shallow. Thanks Ray Alez for the confirmation biased article.
And now with the data that google had offered, it helps me demonstrate this theory. Look at the image and then scroll down for a pseudo-explanation.

Image for post
Image for post

The majority of the posts that got a big percentage of the total number of reactions the reaction wow, had a magnitude of the text lower then 1.

So we have 2 possible explanations. One, the posts that get a big percent of overall reactions as wow does not convey any emotional content. (shallow)

Second, the posts that get a big percent of overall reactions as wow are short in text.

The magnitude of a document’s sentiment indicates how much emotional content is present within the document, and this value is often proportional to the length of the document.

A document with a neutral score (around 0.0) may indicate a low-emotion document, or may indicate mixed emotions, with both high positive and negative values which cancel each out.

This is a list with posts that got at least 1000 reactions, and between 75 and 100% of that reactions were WOW

Image for post
Image for post

A distribution per FB page looks like this. I love it. Using WOW`s seems like a good way to find “sensational posts” posted by the alternative media or other categories.

Image for post
Image for post

What makes a user sad ? Facebook Sad Reactions.

Image for post
Image for post

The top pages that generate the sad emotion are :

Image for post
Image for post

Posts were 80 to 100% of the emotions are sad.

- The Detroit Free Press lost a beloved family member today. Former restaurant critic Sylvia Rector died at the age of 66. She will be missed, and remembered fondly.

- Alan Thicke, best known as the dad on ‘Growing Pains’, dies at 69 after reportedly suffering fatal heart attack.

- A tribute to the iconic world leaders, athletes and celebrities who passed away in 2016.

Almost all articles are related with the death of somebody.

The interesting thing here is combining the information's from the percent of people that are sad to the percent of people that are angry.

If somebody dies because he was old, people will give a sad reaction, leave a comment and they are done.

But if somebody dies because a bomb exploded, then you have a mix of sad and angry reactions, with some weird wow in the middle.

Posts were 60 to 80% of the emotions are sad and 20 to 40% of the emotions are angry.

Here we can mixed posts like

- “I glanced down…and saw my 12-week old fetus. Its head was smashed, and it’s eyes were on either side of its head. Its body was attached to the head, and only the right arm and hand remained. I was shocked and horrified by what I saw.” — source facebook.com/liveaction

- Investigators are trying to piece together the final moments of three children who were killed when they arrived home with their mother to find a gunman waiting inside their Albuquerque, New Mexico home, officials said. — FoxNews

The chilling words … Hundreds of Aleppo men are reported “missing.” Reminds me of Srebrenica. Turned out 8,000 Muslim men and boys who had gone “missing” were in fact slaughtered. —

Posts were the number of sad and angry reactions are almost the same.

- At least 9 people are dead after truck slams into a Christmas market in Berlin and the attacker is on the run. — BretBaierSR

- We can’t forget about Flint.
2017 will mark the three-year anniversary of the water crisis that poisoned thousands of residents in Flint. Still now, many of the city’s residents are still using bottled water to drink, to cook, to bathe.

This Sunday, watch a special 17-minute documentary video on life in Flint created by Ryan Garza and Brian Kaufman of the Detroit Free Press. — facebook.com/detroitfreepress/

I`m curious if we can see a connection between the number of sad vs angry percentage for parts of the world. We take 1000 articles were somebody was killed in the US, 1000 were somebody was killed in Europe, 1000 were somebody was killed in the middle east and we compare the reactions percentage.

Do you Haha ? — Facebook Hahas reactions

Image for post
Image for post

This following pages seems specialized in creating Haha moments with their Facebook users.

Image for post
Image for post

Posts that have at more then 80% Hahas

- Here’s a video that puts the “choke” in artichoke. — FoxNews

- A lone protester outside the Trump victory tour embarrasses herself and threatens to sue Infowars for lawfully filming in a public place. — AlexanderEmerickJones

- Trump is a living parody of himself. This is remarkably ridiculous. — usuncut

- It’s fair to say that Trump supporters aren’t exactly the brightest Cran in the box… — OccupyDemocrats

Posts that have between 60–80% Hahas and more then 10% Angry.

I would rather ________ than read the leftist Washington Post. — Breitbart

Andrea Mitchell’s anti-Trump fretting has officially reached fever pitch! Now she’s hand-wringing that Obama’s climate agenda might dare be called into question under Trump. — newsbusters

Whoopi Goldberg slams Trump! — Breitbart

During his last press conference of 2016, President Barack Obama shared his thoughts on Hillary Clinton’s treatment in this year’s election.
Do you agree or disagree with his statement? — FoxNews

About Me

In the last 3 years i collaborate with the Rise Project, par of the investigative Journalism network, were i do data analysis and pattern recognition to uncover patterns of corruption in unstructured datasets.

In September 2016 i have moved to San Francisco, . Searching for a Job were i can apply my expertise and pay the rent in SFO.

Currently Building a tool that detects .

You can find me online on Medium Florin Badita, , , L O, G Q F

Sometimes i write on my blog

Google Cloud - Community

Google Cloud community articles and blogs

Florin Badita

Written by

Tedx Speaker, Forbes 30 under 30 Europe, European Personality of the Year 2018. Data geek, activist, social entrepreneur. I think all the time about efficiency.

Google Cloud - Community

A collection of technical articles and blogs published or curated by Google Cloud Developer Advocates. The views expressed are those of the authors and don't necessarily reflect those of Google.

Florin Badita

Written by

Tedx Speaker, Forbes 30 under 30 Europe, European Personality of the Year 2018. Data geek, activist, social entrepreneur. I think all the time about efficiency.

Google Cloud - Community

A collection of technical articles and blogs published or curated by Google Cloud Developer Advocates. The views expressed are those of the authors and don't necessarily reflect those of Google.

Medium is an open platform where 170 million readers come to find insightful and dynamic thinking. Here, expert and undiscovered voices alike dive into the heart of any topic and bring new ideas to the surface.

Follow the writers, publications, and topics that matter to you, and you’ll see them on your homepage and in your inbox.

If you have a story to tell, knowledge to share, or a perspective to offer — welcome home. It’s easy and free to post your thinking on any topic.

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store