The anatomy of a hashtag — a visual analysis of the MeToo Movement

Valentina D'Efilippo
Muzli - Design Inspiration
9 min readSep 27, 2018

--

About a year ago, revelation of the scope of Weinstein’s wrongful actions spurred the MeToo movement that saw women worldwide speak up against sexual harassment and assault.

Women’s March attendees in Washington, DC. Photo ©: Shannon Stapleton/Reuters

MeToomentum is a self-initiated data visualization project exploring themes, geographical footprint and key moments of #MeToo.

In a world where sexual abuse and workplace harassment are too common, this online dialogue has become a powerful international movement shedding light on the magnitude of the problem, raising awareness about women’s issues and advancing human rights.

MeToomentum exhibited in London as part of the Data Obscura exhibition (May 2018)

Background

Earlier this year, the Signal Noise team began developing plans for our Data Obscura exhibition, our contribution to The Economist Group’s Open Future Initiative.

We explored how much the information we consume is filtered. Filter bubbles, algorithms, fake news, or just plain old-fashioned ignored counter-arguments. These factors can refract, obscure, and reinforce what we see, what we hear, and — ultimately — what we believe.

As our research developed, I began to fret that we might only capture the dark side of information in the internet age. Certainly, there is value in (and a need for) recognising the challenge of overly-filtered information consumption. But I believe the internet offers us much more than fake news when we question how it is redefining our perceptions of truth.

Social media connects millions of people, and we are spending more of our time on these platforms every day. Social media is an immensely valuable communication tool, which enables everyday people to share and communicate, to celebrate and challenge.

The genesis of this project was a rather simple question: can social media become a vehicle to foster social change, and help reshape traditional views?

MeToomentum addresses this question by dissecting the popular hashtag #MeToo. Focusing on the movement’s first six months, this analysis dissects around 200,000 tweets.

The ambition

In approaching a polarizing subject — especially one which is both personally and societally compelling — it is key that creative energy respects the sober reality of limitations. We faced a number of natural constraints (time, for example) and other constraints more specific to the project, like the completeness of our dataset and the real-time evolution of a cultural movement.

This project has never been intended for scientific research, nor for accurate data analysis. We did not set out to quantify drivers of the MeToo Movement or identify the role it was playing in the related 2017 upheaval. Rather, our ambition for this visual exploration was to capture the shape of the movement’s voices and the impression of these shared experiences.

We wanted to show that our online conversations do not have to sink into oblivion, and that our social media activity represents more than scrolling and liking.

MeToomentum has become a type of archive for the momentum gained by the MeToo Movement. It wishes to celebrate the bravery of all the people who have spoken up, shared their very personal and difficult experiences to provide the possibility for change.

The process

For this investigation, I joined forces with colleague and kickass developer, Lucia Kocincova.

We accessed Twitter API to collect our dataset. From the beginning of October until the end of March, we gathered a sample of tweets containing or related (responses and retweets) to the hashtag MeToo.

It’s worth mentioning that our sample of 200K tweets is a drop in the ocean compared to the millions of posts linked to the Movement. According to the Washington Post, more than 96 million tweets from 2010 to 2017 had something to do with the sexual harassment conversation.

Analysis and processing of the data was a lengthy endeavour, which we approached with careful consideration. Given the magnitude of the dataset, it was impossible to visualize all individual tweets. And, frankly, simply approaching the dataset was daunting.

Thousands of rows of data — recording timestamps and attributes of an online event. These tweets described the disturbing prevalence and harrowing implications of societal flaws. These were personal testimonies of the terror and frequency of sexual violence. Proclamations of solidarity against the imbalance in sexual culture, and courageous messages to victims and perpetrators alike. This problem is real, it’s happened to me, too.

The content was a constant reminder of the gravity of the topic. And emboldened our determination to create something that could faithfully describe the story of thousands of voices.

To find this story, we relied on our curiosity and question-led investigation:

  • What patterns do we see, where are the outliers, what are the key moments?
  • What do the tweets say, are there key emerging themes?
  • Who is tweeting, and who appear to be the primary influencers?
  • What does this look like around the globe?
  • Is this only an American and European experience?
Initial sketches of encoding visual systems

Sketching became a central step between processing the data and a visual system to plot our findings. When it came to outlining the visual structure, we were seeking a metaphor to reinforce some of the movements key characteristics — its complexity, its ability to spread and take hold, and its representation of both strength and fragility.

Semiotic considerations led us to the dandelion as a visual metaphor. Often considered a weed — and certainly a pest to anyone pursuing the perfect green patch — the dandelion is renowned for its ability to spread and grow anywhere. But from another angle, the dandelion is a beautiful and complex flower. One which has been used in popular culture to symbolize endurance, growth, and the possibility of change.

Moodboard that guided me through the process of encoding and designing. The pink used on the website for all the call to actions reference the de-facto symbol of women’s march in 2017 and 2018, the pink pussyhat.

Rooting: what themes emerged?

Finding its roots in historical, systemic gender imbalance, the #MeToo Movement has shed light on the magnitude of the problem of sexual assault and workplace harassment. In this Twitter ‘tsunami’ fuelled by an outpouring of personal perspective, stories shared ranged from tragedy and accusation, to solidarity and calls to action.

To offer a summary of what themes were contained in our dataset, we extracted the top 100 most frequent individual terms and the top 100 most frequent combination of terms (i.e. ‘Sexual harassment’ vs ‘sexual’ on its own).

Many of the words that appear will resonate. From personal names (usually male names emerge as the perpetrators and female names as the victims), to where the violence takes place (‘domestic violence’, ‘red carpet’, ‘sport’, etc.).

We clustered and coloured terms by themes. In the centre, we see violent related words. On the right, some more positive terminology indicating hope/change. While on the top, there are the most frequent individuals named. It’s not surprising that most female names are the victims, and most male names are the perpetrators. ‘Weinstein’ is large, but what is really striking among the male cluster is ‘Trump’ — as frequently mentioned as ‘sexual assault’.

Take a look and see what you can find.

Trending: who are the influencers?

The activist Tarana Burke started the MeToo Movement more than ten years ago, but it gained momentum in October 2017 when celebrity Alyssa Milano shared a viral tweet encouraging victims of sexual assault and harassment to speak out. This is the moment that sparked a global conversation and the rise of the hashtag #MeToo.

Another way to look at our dataset was to identify important moments during these six months. To trace back the dynamic of how the Movement spreads, the key moments that can contextualise the numbers, and the most popular tweeter involved in the conversations.

Although the absolute number of tweets will not account for all of them (due to the limited data sample), we are able to see the overall trend. MeToo has been a powerful enough movement to manifest itself not only online, but also in marches and protests.

We then decided to plot only the popular tweets over time to see which ones received most attention — likes, comments and retweets. Each ‘dandelion seed’ represents a tweet, positioned relative to the date on which it was shared. Its distance from the centre illustrates the number of times it was retweeted, while its size indicates the number of followers the tweeter had.

And this is the result, which is also available online as interactive dataviz:

While media outlets generated significant awareness on Twitter regarding the issue, the hashtag #MeToo has allowed both celebrities and regular people to share their personal feelings and experiences — using social media as a platform for open discussion.

Surprisingly, the trending seeds that travel the furthest (outliers with more that 50K retweets) are a mix of normal people and celebrities.

Spreading: where did the #MeToo Movement spread?

#MeToo has enabled survivors and advocates to overcome geographical distances — and it provides hope that underrepresented voices can help bring about institutional and cultural change.

What started as a movement in North America, in English, has spread to other languages in a significant way. Although English dominates — no surprise as we only scraped the Twitter API containing the English hashtag — several other languages have thousands of related tweets: Spanish, German and French tweets appear in the top 3 in the six months analysed.

We then explored the distribution of tweets across regions and their popularity. The third dandelion of the series shows the number of tweets in each country, but also their reach (orange branches).

Interestingly the movement, which reached every region, was highly popular in countries like Bangladesh, Japan, Turkmenistan, Ethiopia and Serbia.

As Louis Burke mentioned ‘the greatest measurable impact of the MeToo Movement has been in its arrival in countries where ideology or religious doctrine have traditionally hampered open conversations about sexual assault.’

Closing thoughts

Given the attention the project received during the Data Obscura exhibition, we are pleased with the outcome. We hope people will appreciate that we’re trying to give shape to a social media phenomenon . One that represents a defining moment in our society.

However, we don’t think MeToomentum offers a conclusion. The MeToo Movement is just beginning, and its role in reshaping societal change has yet to fully play out. But considering information consumption — and the ways media can obscure what and how we see things — this project illustrates that social media can play a critical role in developing tools that drive social change.

I believe the data visualization can help us see the stories hidden in the data, and also make people sense the implication of these relationships. Rather than being the end of an investigation, it can be the beginning of a conversation. It can be a powerful tool to stimulate dialogue.

So, we would love to hear your thoughts and feedback. Feel free to leave a comment below, contact me on Twitter or email me: hello@valentinadefilippo.co.uk

Special thanks

  • to my collaborator and kickass developer Lucia Kocincova;
  • to Timothy Sondreal and Ian Bamford for their precious help with copy editing and Lukas Kalcok for audio editing;
  • to Erin Gallagher for kindly sharing the Twitter dataset she collected in October 2017 for her stunning network visualization;
  • to our colleagues and friends at Signal Noise for their support and for exhibiting the project in their Data Obscura exhibition;
  • and a very special thanks to all the brave people who have come forward and are making change possible and for the MeToo Movement to support survivors and fight to end sexual violence .

--

--