Dataviz: Samba Charts guidelines

The Challenge

Data is the lifeblood of Samba TV. How do you make Samba TV’s data feel special?

As Samba’s TV data was identified as a competitive advantage, Samba TV needed to make what makes it special tangible and understandable to the market. To do so, we identified four ways to tackle that challenge:

- Raise the level of Samba’s data visualization to the top of the industry

- Standardize how we display data across mediums

- Provide the company with the tools to select and produce the right data viz regardless of skillset

- Push forward new and exciting ways of representing data and insights unique to Samba

‍

Discovery

Dataviz is a little bit like copywriting: everyone is a little familiar with charts, the same way everyone is familiar with writing. Our objective was not to just build charts though: we wanted those charts to tell stories, uncover insights, hit people like a lightbulb breaking. So we needed first to understand what makes it work.

We immersed ourselves in dataviz two ways. First, we wanted to get a headstart by reading some of the biggest dataviz reference books in the space: Dona Wong, Cole Nussbaumer Knaflic, Nathan Yau, Andy Kirk, Scott Berinato: dataviz experts who spent decades thinking and applying their wealth of knowledge in the biggest organizations: Bloomberg, The Wall Street Journal, FiveThirtyEight, Harvard Business Review... We made reading several of their books mandatory for new design hires, so every designer would have a solid foundation on data viz.

The second thing we did was studying years of data viz output from the organizations that are leading the way: insititutions like the Wall Street Journal, Harvard Business Review, but also new forms of journalisms like Axios, Bloomberg Digital, FiveThirtyEight, The Washington Post.

‍

Once we digested this newfound wealth of dataviz knowledge, we started reconciling and synthesizing what we admired and found effective in other orgs’ data viz practices and adapted that into the context of Samba.

We developed the first version of Samba TV’s Data Viz guidelines with the following requirements:

- The guidelines needed to explain how charts would work on multiple formats/medium: presentation decks, prints, web, apps.
- The guidelines needed to help users not only visualize data, but also tell stories, highlight certain insights, and help pick the best chart for the job.
- We needed to anticipate the wide array of contexts the charts were going to be used for: demographics, maps, sports, politics, etc

The guidelines audience were going to be data analysts, research managers, engineers: with them we needed to be as detailed as possible with the details and the execution of these charts.

‍

Every type of chart had specific guidance on when and when not to use it, as well as visual guidelines both for analysts using Pow erpoint or Python, or engineers adapting charts in Javascript.

We also wanted to provide guidance on worst practices, and what to do instead, so we could help people unlearn bad habits, understand why they did them a disservice, and figure out the better solution.

The main thing that we wanted people to remember was that charts were just here as support: what mattered most was the story/the insight they wanted to explain. To help them get there, we standardized the insight presentation around these five elements: the headline, the insight, the key, the chart and the source, giving them equal thought and value. All Samba TV reports adhere to this system.

‍

To ensure we would be able to face any visualization we needed to tackle, we supplemented our brand palette with an extended data viz color palette, shared with our product design system.

Frequent use cases, such as politics, ethinicity or viewership variation were given specific palettes to help people recognize them faster over time.

Diversity and representation are important to our audience, and it was important for us to build a palette that did not alienate them. We steered away from classic, prejudiced color choices and consulted with our audiences to find colors they would be proud of. The dark purple we chose for women references the Suffragettes movement; the yellow for non-binary people is taken from the Non-binary flag

‍

Setting the rules and guidelines were important: the most important part was giving people the ability to easily create charts that would follow these guidelines.

We approached the challenge in three different ways: first, we distributed a 100+ slide deck containing 99% of the charts people would need in their day-to-day.

Second, we worked with our engineering team to standardize charts in our products using React by customizing the Nivo js. library to our guidelines.

Third, we updated the design hub with the guidelines and the templates for easy access.

‍

Samba’s first dataviz guidelines paid off in so many ways: most importantly, they allowed us to confidently step in the world of dataviz journalism.

In the last few years, Samba TV covered important events through the lens of television, such as the 2020 election, the Super Bowl, the Olympics, the rise of streaming, the Marvel Cinematic Universe, diversity and representation on the small screen.

Our charts and infographics made our data more credible, more tangible, and more accessible: since then, our data has been published in AP, Reuters, Bloomberg, The Wall Street Journal, Forbes, Variety, The Hollywood Reporter, Vulture, Adweek. Samba is now a credible source for TV insights.

‍

With newfound strength and confidence in the dataviz space, we brought everything we learned to our research reports, one of the fastest growing revenue lines of Samba.

We redefined Samba’s three major report types with updated charts and layouts, found new ways to make data and findings stand out, and delivered upgraded templates to the Research & Insights team.

In particular, we designed specifically these layouts to push research analysts and managers to think about the insights and the story they were pushing within the slide. We also drove them to annotate and explain insights directly on the charts. The templates helped reframe analysts' mindset on the value we wanted to provide our client, and how they could make said value shine in the reports.

The relaunched insights reports now help Fortune 500 companies better understand the efficiency of their media investments. Between Fortune 500 companies, and the biggest industry trades, Samba TV data viz is closer than ever to its goal of becoming ubiquitous.

‍

Julien has helped us make data science more exciting for us at Samba. During my time as VP of Data Science, he was vital in helping us popularize data science at Samba by making it simple to understand. We worked together on making complex processes and concepts easy to grasp; while doing so, Julien has always been patient and kind, making sure he understood exactly what I was trying to convey and coming up with brilliant solutions. What he does with data is magical.

Maziar Motahari

Global Head of Data Science & Analytics, Uber

Former Vice-President, Data Analytics, Samba TV

‍

Julien has helped us improve our research reports immensely by standardizing our data viz, building templates, writing our data viz guidelines, and on the other hand, pushing the boundaries of traditional data viz in the industry, finding new ways to represent our data sets, uncovering new insights for our clients. He fully understands our clients’ needs, and his encyclopedic knowledge of data visualization has been the secret weapon of our reports. He is a fantastic partner, and I’m always amazed to see what he’s able to come up with.

Karen Biedermann

Director, Media Insights, Samba TV

‍

I enjoyed working with Julien on our media insights projects. Julien has been so successful at bridging the complexity of our data to our audience and clients. He is a dataviz genius, and from the raw data sheets we send him, he always finds so many brilliant ways to represent data, tell stories, and find new insights. On top of that, he is always available to discuss ideas and improve on current visualizations. We are lucky to have him.