OpenVis Conf is a two-day, single track conference centered around the practice of visualizing data with open source tools on the web. Our conference program merges the latest in visualization research, technology, and practice, constructed by an open-call submission and committee review process. In 2018, we followed the main program by a third day of educational, hands-on workshops.Code of Conduct
Sharing their knowledge in 2018
Speaker Video Links
Truth & Beauty Operator
Moritz Stefaner works as a freelance “Truth & Beauty Operator” on the crossroads of data visualization, user interface design and information aesthetics. After successfully quitting academia on the road to a PhD, he has gathered loads of practical experience in drawing insight and beauty from large data collections over the last 15 years. He works for large NGOs, academic and commercial clients, as well as on self-commissioned projects, ranging from tools to manage passenger loads in the German railway system to analyzing Google Trends data to using food and data sculptures for unique data experiences. His work has been awarded ten Information is Beautiful Awards and has been exhibited all around the world. When he is not hopping around the globe, or recording new episodes of Data Stories, he works and lives on the countryside in the north of Germany with his family, horses, and bees.
Machine Learning for Visualization
Others have shown how data visualization can be used to help us understand Machine Learning algorithms, in this talk we will see how we can use Machine Learning to help us make better data visualizations. Recent advances in ML allow us to find representations of large datasets that map surprisingly well to human intuition. By visualizing these representations we can enable new ways of understanding large datasets. This talk will introduce a powerful combination of ML and datavis techniques on an engaging open dataset: millions of Quick, Draw! drawings, and provide clear starting points for generalization to other domains.
Ian is a prototyper, designer and community organizer focused on data visualization for the web. By day he prototypes data visualizations at Google, working on projects for Cloud and Brain. Some evenings he co-organizes the San Francisco Bay Area d3.js meetup and works on the open source tool Blockbuilder. He often tweets about d3.js, data visualization and Machine Learning.
The F Word: Protect Your Work from Four Hidden Fallacies When Working with Data
Without a deep understanding of statistical theory it’s really easy to make mistakes. The ways to correctly use data is often counter-intuitive and weird. This talk will go through four fallacies that are really easy to get caught in when working with data. I’ll show some real world examples of why a working understanding of these four traps is vital for data viz practitioners and how you can check your work to make sure the data story you’re telling isn’t fiction.
The four fallacies we’ll explore are:
- Ecological Fallacy
- False Causality Fallacy
- Prosecutor’s Fallacy
- and Simpson’s Paradox/Lord’s Fallacy.
The focus of the talk will be on actual examples of what each fallacy looks like, how to check your work for these, and a tiny bit of statistical theory. The talk will emphasize the concepts rather than going through heavy mathematical formulas and will be accessible to all levels of statistical experience.
Heather Krause remains unconvinced. As a data scientist and statistician, she battles against both confusion and certainty. She works on the largest questions facing societies today, working with non-profit and social sector organizations, local and national governments, and data journalists. Her pursuit of clarity and realism pushed her beyond pure analysis to mastering the entire data ecosystem including ground-breaking work in data sourcing, design, communication, and journalism, each incorporating bleeding edge theory and technologies. Never satisfied to accept stale answers or perspectives, her methodology is rigorous and unconstrained by convention.
Xenographics: Why We Should All Be William Playfairs
Ever heard of a race concordance chart? A packed bar chart? A Minkowski diagram? A Muller plot? A Hövmoller diagram, maybe? Or the infamous kebab plot? Unless you are a xenographile, like Maarten, you probably haven’t. All these weird charts ("xenographics") go way beyond bars and pies and can be very effective in getting insights out of complex data. With a lot of examples, Maarten will demonstrate the power of weird charts (but also their weaknesses) and explain some of the techniques used to make them.
Maarten is a data journalist and visualisation consultant. A tortuous carreer path brought him from graduating as a bio-engineer in forestry and nature conservation to working as a data journalist at Belgian newspaper De Tijd. He is now working as a freelancer: he designs and develops visualisations, helps to build visualisation tools and gives trainings and workshops. Check his portfolio on his website or follow him on twitter.
University of Edinburgh
Drawing into the AR-Canvas: Designing Data Visualizations for Augmented Reality
Novel display technologies such as Microsoft Hololens, HTC Vive or Oculus have ignited new excitement in the data visualization, VR, and HCI communities. Rather than being concerned about the engineering, we can now concentrate on the design of applications. With respect to data and visualization this means, how can this recent technology help humans to better understand data and the phenomena they represent? This talk summarizes discussions and research on how to think about designing visualization for scenarios in Augmented reality (AR); embedding data visualization into the physical world that surrounds us has recently come into focus as one class of visualization with potentially broad applications.
Lecturer in Design Informatics and Visualization, University of Edinburgh, School of Informatics
Amanda Cox and Kevin Quealy
New York Times
The joke was always that Chapter 7 of our career memoir would be called "disagreements," for all the fights we've had over the years. For a decade now, we've been keeping notes. It's time to resolve a few of them, before that chapter becomes Volume 2 – on stage, together, for the first time.
- Are bar charts really evidence that you live a joyless existence?
- Is good writing more important than compelling charts for persuasive argument?
- Even though we're 99% the same person, why are the charts Kevin is proudest of his simplest, while Amanda remains most intrigued by charts so weird they don't have names?
These are real fights, on which we have real opinions. There will be no fake agreement, no "it depends" mush. We will take sides, and we hope people will leave having chosen one of them. And we have a lot to talk about: We've been working together for 10 years and arguing for about 9 (Kevin was polite for a year).
99% the same person, they are Editor and Deputy Editor of the New York Times' The Upshot.
University of Utah
Designing Effective Visualizations
Designing visualizations requires a wide range of skills, from data wrangling to choosing visual representations to coding. But one skill in particular is essential for designing effective visualizations, and that is figuring out what tasks, and what aspects of the data, you are designing for. The process of going from a high-level, semantically rich question to a set of actionable, data-specific tasks is a messy, iterative process of talking with stakeholders, playing with the data, and exploring design ideas—in this talk I’ll present a concrete framework for navigating the mess. The framework includes guidance for operationalizing questions into tasks, engaging with stakeholders, and using visualizations early and often. I’ll present the framework in the context of successful projects designed with experts in fields ranging from biology to poetry.
Miriah Meyer is an associate professor in the School of Computing at the University of Utah, where she co-directs the Visualization Design Lab. Her research focuses on the design of visualization systems for helping analysts make sense of complex data, as well on the development of design methods for helping visualization designers make sense of real-world problems. Miriah was named a TED fellow and a PopTech science fellow, and has been included on MIT Technology Review's TR35 list of the top young innovators.
Jan Willem Tulp
What I Learned Creating a 3D visualization with 60K+ Datapoints in the Browser
Creating 3D visualizations using WebGL in the browser requires some other strategies than for instance creating 2D SVG visualizations with D3. Using several projects as examples, I will explain some of the technical and design choices that I've made. For instance, what parts are run in shaders, and what parts are not? Or, how to achieve crisp looking texts? And, when combining with D3, what parts are made in D3 and what parts in WebGL, and why?
Jan Willem Tulp is a independent Data Experience Designer from The Netherlands. With his company TULP interactive he creates custom data visualizations for a variety of clients. For more than 6 years he has helped clients such as Google, the European Space Agency, Scientific American, Nature, and the World Economic Forum by creating visualizations, both interactive and in print. His work has appeared in several books and magazines and he speaks regularly at international conferences.
City University of London
Why Not How! Telling Visualization Design Stories
As a community of data visualization practitioners and academics we’ve become good at telling each other what we have created and how we did it. We showcase our work; we create ‘how-to’ books and tutorials; we run technology workshops; we publish our code. What we do less often is explain why we design visualizations the way we do. Yet as a community we would benefit immeasurably from a culture where it was the norm to share our datavis design processes quickly and conveniently.
In this talk I will propose a new approach to data visualization design and communication: Literate Visualization (LV). LV is characterised by: (1) a dialogue between people mixing programming instructions (specification) with prose and visualization (narrative), the emphasis being on communicating our visualization design choices with others; (2) living documents where implementation (e.g. code development) and design narrative evolve in parallel as part of an integrated process; (3) multiple branching narratives that enable the documentation of rejected as well as accepted design choices; (4) ease of construction—doing LV should be as easy as conventional visualization but with support for both narrative authoring and coding.
As well as outlining the case for LV, I will demonstrate it with a new open source programming, markup and narrative construction environment: LitVis.
Jo Wood is a Professor of Visual Analytics at the giCentre, City, University of London. He has interests in visualization design, datavis storytelling and geovisualization. He has authored numerous visualization-oriented software packages over the last three decades, the most recent of which is elm-vega.
Adventures in Space and Time: Visualizing Science Stories in 3D
In a series of projects, I used browser-native animated 3D graphics to tell stories about space and dinosaurs. I’ll go behind the scenes of the construction of these projects to highlight the tools, techniques, and lessons learned in rendering photogrammetry and visualizing data in 3D. Sometimes 3D isn't about 3D: I'll also talk about annotation, how using 2D annotation was necessary to give readers crucial, legible context as stories unfold.
Brian Jacobs is a Senior Graphics Editor at National Geographic. He uses code, design, and visualization to express science topics to the public through interactive maps and graphics. He previously worked for the MIT Senseable City Lab in Singapore and was a Knight-Mozilla fellow at ProPublica in New York.
Ever since the completion of her studies, Caroline Goulard has been nurturing a passion for how information can be expressed, shared and understood. In 2010, sensing that the rich data era will transform the way we work, learn and communicate, she co-founded Dataveyes, a studio specialized in Human-Data interactions. Within Dataveyes, she translates data into interactive experiences, in order to reveal new insightful stories, accompany new data uses and understand our environment shaped by data and algorithms.
Ludovic is a senior interactive designer with more than 10 years of experience in the field. Within Dataveyes, he designs interfaces that fill the gap between data abstraction and users needs. At the crossroads of code and design, he develops data visualizations which are both enjoyable and efficient. Previously, he redesigned a social monitoring tool and managed a creative team in an advertising agency as Artistic Director.
Lessons from a Year of Distilling Machine Learning Research
Chris Olah and Shan Carter have spent a year creating and editing Distill, a machine learning research journal that was created to present current, cutting-edge research in a clear, dynamic and vivid way. They founded the journal to try to reduce the cost of interpreting research, to increase legitimacy for non-traditional research artifacts and to increase the transparency of machine learning. After over a year of trying to execute on these ideas, Shan will share what has worked and what hasn't, which should be useful for anyone interested in using data visualization to make ideas easier to understand, especially in machine learning.
University of Michigan
Tidy Data and Bayesian Analysis Make Uncertainty Visualization Fun
Bayesian statistics gives us a principled way to quantify uncertainty. This talk gives a gentle introduction to Bayesian statistics, and shows how it leads us to a simple representation of uncertainty (samples from probability distributions) that can drive easy-to-build and easy-to-understand uncertainty visualizations. Bayesian modeling and end-user uncertainty visualization are a natural pair: Bayesian modeling techniques produce samples from probability distributions that describe the uncertainty in estimates and predictions, and a growing body of research suggests that sampling-based visualizations of uncertainty lead to better estimates and better decisions from users. I will tour a variety of uncertainty visualization techniques and domains (including medical risk communication, hurricane path prediction, and real-time transit arrival prediction) where sampling-based uncertainty visualization has been successfully applied, and show how to generate such visualizations using Bayesian approaches.
Matthew Kay is an Assistant Professor at the University of Michigan School of Information working in human-computer interaction and information visualization. His research areas include uncertainty visualization, personal health informatics, and usable statistical tools and analysis. He is intrigued by domains where complex information (like uncertainty) must be communicated to broad audiences (as in health risks, transit prediction, or weather forecasting).
New York Times
Visualizing Climate Change
Climate change is a complex problem that can feel distant in time and space. At The New York Times, my job is to make sense of the causes and consequences of a warming world using maps, charts, and pictures, as well as words. I’ll talk about some of the goals and the challenges of translating climate science for readers who (by and large) aren’t experts. How do we better help readers understand current and future impacts, both at home and around the world? How can we inform without overwhelming?
Nadja is a journalist at The New York Times, where she makes graphics and writes about climate change and the environment. She previously worked as an interactive editor at the Guardian's US office in New York, where she focused on science, health and politics.
Co-founder and Engineer, Satellite Studio
OpenStreetMap: The Planet is the Dataset!
OpenStreetMap is the map of the world that is made by everyone and that belongs to everyone - aka "the Wikipedia of maps." It is a huge and truly fascinating dataset, whose richness and potential can be overlooked. This talk is meant both as an exploration and as a survival kit for understanding, consuming and manipulating OpenStreetMap data. We will see how the tools created by the OpenStreetMap community can be used to craft maps from this immense dataset. We’ll go from city-scale, creating thematic maps with tools such as Overpass and Carto, to planet-scale, generating vector tiles (tilereduce, tippecanoe, mbtiles) and rendering them with WebGL. The audience should leave with a better idea of what OpenStreetMap data is, how it is relevant for context and thematic mapping, and with a practical understanding of its tools and best practices.
After working at CARTO as a community technologist, then at Vizzuality as a front-end engineer, he recently joined GlobalFishingWatch, an NGO dedicated to monitoring fishing activity globally, and co-founded Satellite Studio, a mapping and visualisation studio creating unconventional interactive experiences on screens and beyond.
Thinking with Data Visualizations, Fast and Slow
Your visual system evolved and developed to process the scenes, faces, and objects of the natural world. Using that system to process the artificial world of data visualizations is an adaptation that can lead to fast and powerful—or slow and inefficient—visual processing, as we will experience through interactive demonstrations of new research findings on visual capacity limits. Understanding these limits produces guidelines for constructing effective visualizations for both visual analytics and visual communication of patterns in data, and explains how display designs and motivated cognition can bias interpretations of those patterns.
Steven Franconeri is a Professor of Psychology at Northwestern University, and Director of the Northwestern Cognitive Science Program. His research is on visual thinking, visual communication, and the psychology of data visualization. Northwestern's Visual Thinking Laboratory explores the power and limits of your visual system, and how better design and pedagogy can help students and scientists understand and use visual representations across paper, screens, and the imagination.
Visual Narratives for Empathy, Emotions, and Comprehension
The talk will focus on the design process behind a set of data visualization and data art projects, characterized by strong differences in terms of represented information and usage context, but with a common ground of interest for the human relationship between designer and user and for a constant visual experimentation. The projects will be explored through a presentation of the design phases — from data to visualizations passing through visual inspiration — and of the lessons learnt throughout the processes, with a reflection on how giving shape to data can become a generator of empathy, emotions, and comprehension.
Federica Fragapane is an award-winning freelance information designer. Among her projects, she has designed data visualizations for the United Nations Environment Programme and WIRED magazine. She collaborates with the Italian newspaper Il Corriere della Sera, working on the analysis and visualization of cultural, environmental and social topics for the cultural supplement "La Lettura."
She is co-author of Pianeta Terra, an infographics children's book published by National Geographic Kids and White Star Kids and author of The Stories Behind a Line, a visual narrative of six asylum seekers’ journeys. She strongly believes in the communicative potential of data visualization and she's constantly testing it. She can be found at @fedfragapane.
The Washington Post
How Data, and the Visualization of It, Helps Us Understand "Us"
For the Paris World Fair in 1900, African American philosopher and teacher W.E.B. Du Bois commissioned several data visualizations from Atlanta University that detailed the education, income, vocation and residence of African Americans in the United States. He did this just 35 years after the end of slavery.
As the century advanced, the U.S. continued to collect more data on race through the U.S. Census and American Community Survey. This talk will explore the history of race data in the United States and abroad. We'll discuss how this data has been visualized and it's impact on research and storytelling.
Aaron Williams is a reporter who specializes in data analysis and visualization for The Washington Post. He’s also a co-founder of the California Civic Data Coalition and on the advisory board for Open News. He previously worked for the San Francisco Chronicle, Center for Investigative Reporting and Los Angeles Times.
Data Visualization Freelancer
"Geogiffery" and Other Experimental Ways to Visualize Geospatial Data
Animated maps are an excellent way to display spatial data. They are a very efficient way to display movement data, showing things in sequence or highlighting the geographic spread of something over time. Often a low-tech "geogif" is only an entry point to data and will not allow users to delve more deeply, but in many cases it can be a good way to highlight the big picture. In my talk I will present the main tools (PostGIS, QGIS) I use and open up the process of making simple geospatial animations. I would also like to show how other experimental ways of visualizing geospatial data can stretch the limits between normal data visualization, map making and even data art.
Topi Tjukanov is a geospatial data enthusiast and a data visualization freelancer. You can view his work on his site.
How Games Use Dataviz to Guide Behavior
Inside any game, how and why any player makes a decision is usually crucially supported by visualizations of data in the game. Dataviz in these scenarios is constantly adapting to the player’s actions, warning them of danger, and rewarding them for good behavior. While a lot of dataviz work in publications and in academia help to explain something or provide insights, the visualizations that already exist in games are constantly trying to guide change — something we can all learn (and steal) from in our work. I’ll go over the most popular game in the world right now, League of Legends (100 million players each month) and long-time popular games like Mario Kart, and explain how each use dataviz in ways to provide feedback loops and affect behavior.
Sisi Wei is the deputy editor for news applications at ProPublica, where she edits a team of investigative journalists/developers who build interactive stories to serve the public interest. Sisi's work has ranged from investigating which U.S. colleges saddle students with debt to monitoring how often China blocks international news outlets. Sisi has won numerous Malofiej, SND Digital and ONA awards, and she is also the co-founder of "Code with me," a nonprofit that teaches journalists how to code. She previously worked at the Washington Post, the Wall Street Journal and the Associated Press.
The Data R&D Institute at emlyon business school in Lyon, France, is a proud primary sponsor of the conference in 2018. We are a growing group of data science instructors and developers bringing data literacy to a changing business climate. The conference is a highlight for the school, allowing emlyon to showcase leadership in the fast-paced, technical landscape of businesses centered around data science and data design.
Qlik Playground is a free programming environment where you can learn about, use and experiment with Qlik’s Associative Engine and APIs. By accessing public data sets (including their own GitHub data) through our APIs, you can create cool data-driven apps using powerful search, data retrieval without writing queries, charts with a single line of code and more.
Qlik Branch is a free community of more than 20,000 developers where you can share and collaborate on open-source projects and ignite innovation leveraging Qlik products. Here you can share, download and give feedback on projects that plug into Qlik products; connect in real-time via Slack; and learn via the blog and upcoming Knowledge Hub.
Datadog is a SaaS-based monitoring and analytics platform for large-scale applications and infrastructure. Combining real-time logs, metrics from servers, containers, databases, and applications with end-to-end tracing, Datadog delivers actionable alerts and powerful visualizations to provide full-stack observability. Datadog includes over 200 vendor-supported integrations and APM libraries for several languages. Find out more at on Twitter at @datadoghq, on Facebook, on Linked In, and check their Careers page!
OpenVis Conf relies heavily on sponsors to allow us to put together the conference our attendees deserve and expect. There are many ways in which you can help: From sponsoring a diversity scholarship to throwing a party for our attendees. We're adding a job poster board this year, too. Let us work with you. Find out more in our prospectus, and get in touch!Download Prospectus Contact Us
Curating the conference
The Program Committee solicits and reviews all the talk proposals. They do an incredibly difficult job under a tight deadline, using their own donated time. Representing data practitioners, academics, journalists, and teachers, they make OpenVis Conf an exceptional community-driven event.