The year is not over yet, but if we had to choose which one was the top data visualization event of 2014, only a handful would be even considered – and Bocoup OpenVis Conference would be on that list. The two-day event took place in April, with keynotes from the likes of Mike Bostock, Eric Fischer, Andy Kirk and Robert Simmon, among others. You can catch up with all the presentations here.
The third OpenVis Conf will take place on April 6th & 7th of 2015. The announcement came yesterday, and the call for speakers is now open – this year, not only can you submit a talk, but you can also suggest a topic OpenVis should cover or even a specific speaker you’d like to see.
We managed to ask Irene a few questions about OpenVis Conference and all the other great stuff she’s been doing at Bocoup, her career in programming and data visualization, and how she thinks the field will evolve in the next years.
Visual Loop (VL) - Irene, does it make any sense, at this time, to talk about Open Data without associating to it, by default, data visualization?
Irene Ros (IR) – Data without some sort of analysis is just noise. If it’s raining outside right now, we know it’s raining, but that’s it. However, knowing what the weather was the entire week, we can probably make some conclusion about the season we’re in and what kind of weather we’re expecting. This is true about all forms of data, sometimes there’s information in it, and sometimes there isn’t, but data isn’t information in it of itself.
Seeing the results of data analysis visually isn’t a requirement. There are absolutely those select few brilliant statisticians who can look at numeric results and get a reaction to what is happening, but I would say those are few and far in between. We humans have 5 amazing senses and we can only augment our understanding of the world if we get to use them. We are incredibly good at seeing patterns, outliers and trends if given the right forms, which is why there is so much love for data visualization in the media - we are democratizing data analysis by reaching more people using visual methods.
Our data visualization work at Bocoup falls into two camps - building exploratory tools and creating user-facing visualizations. Building either of those often relies on understanding the questions we’re asking, figuring out if they can be answered with the data (an analysis process that often involves “sketch” visualizations of our own for the purposes of understanding our dataset) and designing & building the best visual interfaces for answering those questions, which often involves cycles of prototypes and revisions. For example, if our data has a lot of outliers, and we’re particularly interested in those, we’ll gravitate towards visual methods that give more prominence to those outliers, whereas if we’re interested in overall trends, we might focus on the patterns. There really isn’t a single visual method that is always the right answer, which is really part of the fun.
Finding the right visual method is only half the battle though, and I wish we talked more about the other half. Providing context, showing our own uncertainty, offering some narrative and guidance are all things that are critical to augmenting a user’s experience and understanding. If we want to improve the visual data literacy of our audience, then we are responsible for providing those things. Unfortunately data visualization’s popularity means that you can find it in contexts that use it for manipulation rather than informing the audience, like advertising. To combat that, we should be arming our audience with the knowledge of discerning between what is accurate and truthful and what is otherwise just advertising cleverly disguised.
Data without some sort of analysis is just noise. If it’s raining outside right now, we know it’s raining, but that’s it. However, knowing what the weather was the entire week, we can probably make some conclusion about the season we’re in and what kind of weather we’re expecting.
VL - You’ve been a programmer for over two decades now. In what stage of your learning process you had your first contact with data visualization?
IR – You are betraying my age! I started programming when I was 9 and haven’t really stopped since, although my role has evolved greatly in the last few years. I am a very visual person and always struggled to enjoy programming tasks that didn’t result in something visual and engaging. My breakthrough happened during my second year at IBM. In 2007 I saw a talk by Martin Watternberg and Fernanda Viegas who spoke about Many Eyes and suddenly everything clicked. I saw a beautiful conflation of my fascination with visual forms, math (which was my favorite study subject) and programming (which was my favorite tool for the job.) I joined their group working on Many Eyes and have been a data visualization practitioner since.
In retrospect, I really wish I was exposed to human-computer interaction and data visualization as concepts sooner. I saw it in the world, but never distinguished it as its own “thing.” I think if I knew that there was a way to make visual impact with programming, it would have given me a purpose sooner. Whenever I visit my Computer Science department at UMass Amherst, which I love dearly, I make a point of exposing other undergraduates to this world.
VL - And can you point out three other defining moments in your career, regarding data visualization?
IR – Certainly joining IBM Research’s Visual Communication Lab was a defining moment. While I had conducted research as an undergraduate, my exposure to real world paper authoring, user studies, the review process and the general academic community were all new. It wasn’t until a few years ago that I truly appreciated just how much thought we put into projects, prototypes, drafts etc. I am eternally grateful to Martin who taught me to care about my output and not my code, and to throw things away and start over when they aren’t working.
A project I think about a lot is Many Bills, on which I worked during my time at VCL with Yannick Assogba. I was very interested in government transparency and we as a group were exploring text analysis. We built a really neat web application for visualizing congressional legislation. While technically we built something really cool and beautiful, we didn’t get the audience engagement we wanted. As it turned out, people weren’t that interested in understanding federal legislation in general. It taught me to think about my users first, and my visualizations second. I am not ashamed to use a simple chart if that’s the best tool for the job, because fundamentally my goal is always to convey some information in an engaging and informative way, not to just create a visualization.
The third most defining moment was obviously joining Bocoup. I really enjoyed being a part of the academic community but really wanted to put my learning into practice and to support others like me, who were stumbling into the field from other disciplines and were finding it hard to build up their knowledge and practice. Our work with the Guardian Interactive team taught me a lot about journalistic and analytical integrity and laid the foundation for the Miso Project and the rest of our data visualization practice.
These days I try to think about impact a lot and carry with me a sense of responsibility for others (practitioners and audience) in my work.
Finding the right visual method is only half the battle though, and I wish we talked more about the other half. Providing context, showing our own uncertainty, offering some narrative and guidance are all things that are critical to augmenting a user’s experience and understanding. If we want to improve the visual data literacy of our audience, then we are responsible for providing those things.
VL – Now, at Bocoup, a lot of people recognize you by the tremendous success the OpenVis Conference achieved in the first two editions. But of course that your work over there goes way beyond that conference, right?
IR – It sure does. We’ve been doing a lot of data visualization related work here at Bocoup.
Our first overlap between data visualization and the open source world was with the Miso Project. It was during my time at the Guardian when we got to experience the lack of tooling for building web based data driven narratives. We started working on the Dataset library then, followed up by Storyboard (a controlflow promisesbased library) and continuing more recently with d3.chart (a wrapper around d3.js that allows one to create reusable charts.) We are continuing to grow those libraries and are always looking for challenges our community is facing and trying to come up with solutions we can create and share with others. The Miso Project has seen some incredible uses over the years and we’re very committed to its success.
We are always thinking about education at Bocoup. I’ve recently started teaching a d3.js class which has been a very rewarding experience. It’s such a powerful and complicated library! I greatly appreciate empowering our students to create meaningful visualizations after only 2 days in class. I’ve been really impressed with what they’ve been able to do since. We’re always thinking hard about our curriculum.
Of course a lot of our work happens with our clients. We have an incredible range of projects going at any time from very creative to very analytical (and both!) We work with folks along the full spectrum of creating datadriven interfaces and stories from data analysis to design & prototyping to implementation and deployment. More recently, one of my favorite projects was working with Climate Central to create an exploratory tool demonstrating the risks of rising sea levels. Not only did we have some interesting design & data challenges, but we had a greater sense of purpose in our work. Being able to convert data into actionable and locally relevant information was a really rewarding experience.
The past year has been my first as the Data Visualization Practice Lead here at Bocoup. I get to come to work everyday with the purpose of helping and enabling my teammates, clients and our community to become better data visualization practitioners and consumers. It’s a great combination of allowing me to think critically about our field while working in it and towards improving it going forward.
VL - Like we said, this last edition of OpenVis Conference was a huge success. How challenging and rewarding was it?
IR – It really was an incredible experience. I can’t say enough about how rewarding it was to create an event that so many people found useful and informative. It was my dream to create a conference that catered to those who were learning & improving, as well as those who were leaders in our field and wanted to share their work. It’s not easy bringing those two communities together sometimes, but we’re very dedicated to doing that. For those that haven’t seen them yet, the videos for OpenVis are posted on the website and are absolutely worth a watch.
It was most certainly a challenging endeavour. We knew from the get go that we weren’t going to cut corners or make compromises that would in any way reduce the quality of the program. It meant that sometimes we had to get creative or work extra hard, but the program and our attendees’ experience came first. I will say that the programming aspect of OpenVis has been a very humbling experience. The programming committee and myself dedicated dozens of hours to reviewing and searching for practitioners who we wanted to bring into the spotlight. There are a lot of prominent figures in our field, but there are also incredible people who are doing amazing work and often don’t get the recognition their work deserves. We want to find the best talks, the best projects, the best tools and the best open source solutions to the problems we face and that search is a matter of extreme dedication. I couldn’t be more grateful to the committee for their work.
Another challenge we think about often is that of diversity. Bocoup is an incredible place when it comes to diversity and our conferences and events channel that spirit. It’s so important for us to create a safe environment both in and outside our own events. As a result, we work tirelessly to have as diverse a speaker set and audience as we can. We aren’t always as successful as we’d like (and that alone can be an entire conversation) but we make it a high priority and will continue to do so.
As far as OpenVis Conf 2015, we announced it yesterday. You should all take a look at http://openvisconf.com and submit a talk, recommend a speaker or suggest a topic and get your early bird tickets. I’m really excited to put this year’s program together. Folks should also follow @OpenVisConf for future announcements, it’s going to be a great one!
VL - Great, sound like it’s going to be another huge hit! And speaking of events, we had the pleasure to interview Mariana Santos a couple of months ago, and among other topics we discussed the Chicas Poderosas project, something that you are also involved in someway, right?
IR – That is correct. I met Mariana during my time at the Guardian. She has since done some incredible things, including founding Chicas Poderosas. The goal of the organization is to empower women in newsrooms throughout latin america to work with technology. I joined Mariana for an event in Costa Rica where we ran a two day workshop aimed at teaching data analysis and visualization related skills. The women did some incredible investigative reporting - I actually think a few of those projects were reported in the paper. I was so incredibly impressed and humbled by these women!
I feel very fortunate to work at a company that cares so much about diversity in the programming community. To support the Open Web isn’t just about open source software, but also about creating an open place to share and participate in technology and society. We organize a lot of community events (like the Boston Data Vis meetup) and work with local organizations like the Science Club for Girls to continue making the open web a friendly place for everyone.
VL – In a recent interview, you mentioned you were excited about how far mapping has evolved on the web. How do you see the data visualization field evolve, as a whole, in the next decade? Care to risk any predictions?
IR – That’s a very interesting question and one I think about a lot. I feel a second wave of “make your own data visualization tools” coming. Between Lyra from Stanford and Datapad (which I haven’t seen yet, but I know the team behind it is incredibly capable,) we are going to see a resurgence of webbased analysis and visualization tools. I foresee more tools that bring together analysis and visualization coming to life.
I also see a huge thirst for data visualization in new fields. I think people know that it’s a powerful device are looking to integrate it into their workflow and communication. Front end engineers who traditionally worked within the constraints of CRUD apps are having to also add charting and think more about data driven behavior. This means we will need more high quality educational content and better integration points between traditional web application frameworks and data visualization. It will blur the line, but I think that’s fundamentally a good thing. I’d love to see a world where we don’t separate data visualization from other web content, but create them side by side.
I’d love to see a world where we don’t separate data visualization from other web content, but create them side by side.
VL- And to close, Irene, what can you tell us about new projects? What can we expect from Bocoup for the upcoming months?
IR – Well, there won’t be a dull moment, I can promise you that! We’re working hard on OpenVis Conf 2015 , so stay tuned for that. We are also working on some educational materials that will shed light on some of the trickier parts of d3, so keep an eye on the bocoup blog and online education page. We are working on some exciting projects as well, so we’ll be posting those on our data visualization portfolio.
VL Thank you so much, Irene!
IR – Thank you!
We really appreciate the time Irene dedicated to answer in such depth all our questions. You can keep up with Irene’s updates on her website, and connect with her on Twitter (@ireneros), LinkedIn and GitHub.