Martin Wattenberg — Name Voyager — popular data-visualization breakthrough

2005

Artists: Martin Wattenberg · Year: 2005 · Medium: Interactive baby-name data visualization — stream-graph showing name popularity over time · Substrate: Java applet; Social Security Administration baby-name dataset (1880-2004); stream-graph visualization with type-ahead filtering · Place: Cambridge, Massachusetts, USA · Institution: IBM Visual Communication Lab (later Google) Martin Wattenberg launched Name Voyager in 2005 on the website of Laura Wattenberg's Baby Name Wizard (Laura is Martin's now-former spouse; Martin built the visualization as a supporting tool for her popular name-history book and website). Name Voyager is a real-time stacked-graph visualization of the popularity of every US baby name from 1880 to the present, queryable by typing letters into a search box. Type "A" and the graph displays every name beginning with A and their relative popularity over 130 years. Type "Al" and the graph filters to Alexander, Albert, Alice, and so on. Type the full name and you get that name alone, in full temporal detail. The visualization went viral in the specifically 2005 sense of the term: within a few days of launch, Name Voyager was referenced on the front pages of the New York Times, Slate, and numerous technology blogs; within a month, it had been cited in over a hundred subsequent data-visualization articles; within a year, it had become one of the canonical reference examples for the emerging discipline of information design. What Wattenberg did with Name Voyager is a specific argument about what data visualization can do. Most visualizations of baby-name data at the time (the rare cases where such visualizations existed) presented static charts of the top-ten names in specific years. Wattenberg's innovation was to make the filter itself the interaction: the viewer types letters, and the graph responds instantly, producing a tool that felt more like a musical instrument than like a chart. The underlying dataset was the same one other researchers had been using; the interaction model was what was new, and it was the interaction model that made the piece circulate. Name Voyager directly shaped the subsequent decade of data-visualization journalism. The New York Times interactive department, which hired Aaron Koblin and others from the same 2005 moment, treats Name Voyager as a founding reference. Wattenberg and Fernanda Viégas subsequently built the Many Eyes platform at IBM (2007, elsewhere in this archive), and then moved to Google's PAIR research group, where they continue to work on machine-learning visualization. Lore Wattenberg built Name Voyager over a couple of months in evenings and weekends — the data was the US Social Security Administration's public baby-name dataset, which Laura Wattenberg had already cleaned for her book. The visualization was rendered in Java (which was then still the standard for interactive web graphics). Martin has said, in subsequent interviews, that he was genuinely surprised by how quickly the piece spread; he had expected Laura's existing readers to use it, not a global audience. Within a week of launch, the server was overwhelmed and Laura Wattenberg had to upgrade her hosting three times. The original Java applet stopped running in most browsers by the late 2010s as Java web-applet support was deprecated; a subsequent JavaScript port has kept the visualization alive. Sources • Wattenberg, Martin — "Baby Names, Visualization, and Social Data Analysis" (IEEE InfoVis, 2005). • Baby Name Wizard — babynamewizard.com (Name Voyager site, still active). • Viégas, Fernanda B.; Wattenberg, Martin — bewitched.com (joint archive).