GitHub - jadianes/data-journalism: Data journalism and easy to replicate notebooks using Python, R, and Web visualisations. Disinformation Visualization: How to lie with datavis. By Mushon Zer-Aviv, January 31, 2014 Seeing is believing.
When working with raw data we’re often encouraged to present it differently, to give it a form, to map it or visualize it. But all maps lie. In fact, maps have to lie, otherwise they wouldn't be useful. Some are transparent and obvious lies, such as a tree icon on a map often represents more than one tree. It all sounds very sinister, and indeed sometimes it is. Over the past year I’ve had a few opportunities to run Disinformation Visualization workshops, encouraging activists, designers, statisticians, analysts, researchers, technologists and artists to visualize lies. Centuries before big data, computer graphics and social media collided and gave us the datavis explosion, visualization was mostly a scientific tool for inquiry and documentation. Reproducing Lies Let’s set up some rules. We don’t spread visual lies by presenting false data.
Should we legalize the killing of babies? I would hope most of you would say: No. How to Lie with Data Visualization. Data visualization is one of the most important tools we have to analyze data.
But it’s just as easy to mislead as it is to educate using charts and graphs. In this article we’ll take a look at 3 of the most common ways in which visualizations can be misleading. Truncated Y-Axis One of the easiest ways to misrepresent your data is by messing with the y-axis of a bar graph, line graph, or scatter plot. In most cases, the y-axis ranges from 0 to a maximum value that encompasses the range of the data. Let's see how this works in practice. On the left, we’ve constrained the y-axis to range from 3.140% to 3.154%. SND 2016: Design in the Age of Statistics-Driven Journalism - Google Slides. Some things I learned about data-driven storytelling in Schloss Dagstuhl — Data Driven Storytelling. I learned that much of data storytelling can actually be understood much better when thinking about formats like speeches, presentations, jokes, documentary film, rather than “tales” or “plots”.
Telling a story does not automatically imply a simplistic, author-driven, linear, primarily entertaining narration. Really interesting stuff can happen when we inject mechanisms from these other — persuasive or entertaining — forms of communication of information. Look at the beautiful use of repetition and rhetorical questions in this lovely Bloomberg piece: In its widest form, storytelling is about establishing a flow of data perspectives. Defining and redefining data perspective can become a storytelling mechanism in itself, like in this blog post on biking accidents. Projects like netwars take these ideas even further, and start to blend documentary movie, graphic novels with data visualization and multi-platform experiences, weaving an atmospheric net around a topic. So — watch this space! - Course - LEARNO. Course overview Getting started with Google refinements (27:56)Advanced Image Search (12:02) Module 1: Verification in a networked world by Craig Silverman Introduction (01:21) The Fundamentals (14:50) Open Verification (20:11) Verifying During a Crisis (15:27) Module 2: Using and Verifying User Generated Content by Claire Wardle Introduction (02:21) Key techniques (20:34) Ethics and Protecting your Sources (14:57) Crediting and Labelling (16:55) Module 1: Data journalism in the newsroom by Simon Rogers This module is an introduction to data journalism.
The story of a transformation, in three years. How Julius Troeger, a journalist working for Berliner Morgenpost, managed to create a string of impressive data interactives and won the support of his publisher for a small, but growing interactive team in the newsroom.
If data journalists could write a wish list, it would look like this: to have the time to work on interactive stories - days, even weeks, not just hours, to be supported by a skilled and capable team, encouragement from upper management, and, finally, a budget large enough to achieve all of the above. Julius Troeger, a German journalist working for Berliner Morgenpost, can tick off all of these items. He is heading a small, but growing team, which has already published a great string of award-winning stories. It took him roughly three years. Today, Berliner Morgenpost is known for its outstanding data interactives. Importantly, the new management understood the value of the work done by the interactive team. Interview. The Journalist-Engineer. A couple months ago, I published an article comparing historic and present-day popularity of older music.
I used two huge datasets: 50,000 Billboard songs and 1,4M tracks on Spotify. If I were writing an academic paper, I’d do a ton of analysis, regression, and modeling to figure out why certain songs have become more popular over time. Or I could just make some sick visualizations… Instead of reporting on my “theory”, I wagered that readers would get more out of an elegant presentation of the data, not an analysis of it. It’s a completely different approach to storytelling. Here’s that same approach on another project: rappers and the size of their vocabulary. Instead of proving that one rapper was better than another, readers are really good at absorbing the data, and they’d much rather form their own judgements. A few years ago, Bret Victor wrote about the notion of passive and active readers: In theory, this sounds great…but kinda crazy.
But it’s happening — there are active readers.