The Data Storytelling Workbook

Anna Feigenbaum, Aria Alamalhodaei

• Publisher: Routledge
• Year: 2020
• ISBN:
  9781351684705
• Edition: 1
• Topic: Social Sciences
• Subtopic: Media Studies
• Index: Social Sciences

From tracking down information to symbolising human experiences, this book is your guide to telling more effective, empathetic and evidence-based data stories.

Drawing on cross-disciplinary research and first-hand accounts of projects ranging from public health to housing justice, The Data Storytelling Workbook introduces key concepts, challenges and problem-solving strategies in the emerging field of data storytelling. Filled with practical exercises and activities, the workbook offers interactive training materials that can be used for teaching and professional development. By approaching both 'data' and 'storytelling' in a broad sense, the book combines theory and practice around real-world data storytelling scenarios, offering critical reflection alongside practical and creative solutions to challenges in the data storytelling process, from tracking down hard to find information, to the ethics of visualising difficult subjects like death and human rights.

Visual Data Storytelling

Visual Data Storytelling

During the 19th century visualising information came into what Michael Friendly (2008) calls the ‘Golden Age of Statistical Graphics.’ This accompanied the rise of state-run statistical offices collecting demographic data, alongside the increased use of statistics in the sciences, astronomy, urban planning, and commerce. Quantitative data collection and evaluative thinking gave way to a range of graphical developments for representing statistical information. While government institutions were the primary users of these new graphic techniques, this period also saw the employment of information visualisation for social campaigns, including John Snow’s cholera map and Florence Nightingale’s rose diagram on preventable deaths in the Crimea War.

→ You can learn more about John Snow’s Cholera map in the section The Problem with Maps as Representations and Florence Nightingale’s rose diagram in Defining Big Data.

→ D3 is a JavaScript library for visualising data using web standards. It can help visualisation designers produce dynamic, interactive data visualisations for web browsers. Gephi is a non-profit Open Graph Viz Platform that is widely used by academics and industry for both exploring and communicating data, particularly related to network analysis. Many of the data visualisations you see around the web are made with these popular resources. For more information, go to www­.ge­phi­.or­g and www­.d3­js.­org­.

→ Tableau is primarily designed for data analysts within businesses. However, while it is generally used internally by organisations, Tableau Public has point and click functionality that allows people to generate visualisations without the need for extensive coding. Many journalists, non-profits, political commentators and fan communities use this version of Tableau to easily create and share visualisations. Extensive galleries, online tips, and blog posts are available to explore at www­.ta­ble­au.­com­.

Today, 19th century conventions for the graphical display of information continue to be used by data visualisation designers (Kennedy et al., 2016). Visualisation libraries like D3 and Gephi, as well as software programmes like Tableau, embed these conventions into contemporary design practice. There are many benefits to these platforms in terms of saving time and resources. First, these platforms make it increasingly easy to visualise data without needing to participate in other parts of the data process, i.e. its collection, cleaning, or even analysis. Second, templates are often already set up with ‘best practice’ chart design, colours, and caption placement. This frees users from needing to make these decisions. For audiences, the repeated use of templates can help build literacy through repetition. As they rely on dominant forms of graphical representation that have been in circulation often for over 100 years, the viewer’s eyes are already trained to interpret what they are seeing.

At the same time, templates can also limit creative potential and opportunities for innovative data exploration and storytelling. In addition, because templates repeat dominant representational rules and forms, the content can get diluted by the sameness of the representation. In other words, when the same graphic symbols and colour schemes tell the story of fuel prices, animal shelters, political races, and sports statistics, the individuality of the data and its story can get lost. This template fatigue can lead to people seeing the data point but not connecting with the subjects of data. Too much use of the same template comes with the risk of creating visual data stories that get further and further abstracted from the question of ‘what is at stake?’

Too much plugging data into templates without reflection can also make it difficult to answer the question ‘Why does this visualisation matter?’ When the process of designing a visualisation is detached from the process of data collection, preparation, and analysis, it can be hard to create a meaningful data story. Understanding the context, backstory, or potential biases of the data you are working with is important for adding meaning. Without this understanding, you are left with only the surface to work with.

This is why our BU Civic Media Hub approach aims to be more holistic. Along with many of the other data storytelling teams discussed in this workbook, we integrate visualisation into the process of our data storytelling. As visualisations are both representational artefacts and exploratory tools, they can lead to richer understandings of the data you are working with. Thus, rather than treat visualisation as a separate design step that comes after all the data analysis is complete, it can function as part of a two-way communication system. Taking this approach, we find that often the most important questions emerge during the process of visualising for our data stories: What standpoints are being privileged? Who or what is missing? How can we better account for the biases and backstory in our datasets?

In order for data visualisation to be about more than making information beautiful, it must be approached with the same reflective and critical thinking as any other aspect of a data storytelling project. This is why we believe that people should be equipped with the conceptual and practical tools needed to tell visual data stories that matter. These tools are what help us grapple with the challenges of graphic design and visual storytelling. They are what allow us to engage with the challenges of sensitive subjects, missing data, messy data, data biases, and all the other obstacles that working with data throws our way.

Feminist Data Visualisation

Our approach to visual data storytelling is indebted to recent work in feminist data visualisation. One of the main things that feminist data visualisation is concerned with is how biases in perception and representation can become embedded into design practice. These biases often get repeated in data visualisations without reflection or questioning. As a result, negative stereotypes can be reproduced, marginalised voices can get erased or obscured, and many issues can get left out of visual data stories all together.

→ In visual culture, binaries refer to binary oppositions or two categories that are set in contrast to each other, such as man/woman, love/hate, nature/culture, straight/gay. Bina-ry oppositions are also often linked together, for example in dominant stereotypes of ‘mother earth’: woman-love...