We sought new ways to present Census data on an urgent, but often unseen, issue in the United States: food insecurity. While Census data undoubtedly proves the scope and characteristics of hunger in the United States, it can obscure the lived experience for each one of the 49 million households that are food insecure. Our project aims to surface these stories, combining a data-driven national level picture of food security with the first person accounts of hunger from everyday Americans. By combining the power of storytelling and data, we aim to create a participatory platform where users can submit their own stories, reflections, and reactions on the issue of food insecurity, continuously enriching the content and surfacing new stories.

Working with the US Census Bureau, the Computational Creativity Lab was tasked with using Decennial Census Data to create tools that advance the Bureau’s diversity, equity, and inclusion goals. Speaking with local librarians in Pittsburgh, PA, we learned that efforts to diversify library collections are often stalled by poor data collection. Librarians often have no way of evaluating how diverse their library’s collections are, or where there might be underrepresented authors. To address this need, we built a participatory data collection system where participants are invited to answer a series of questions which evaluate a book along certain diversity criteria. The outcome is a visualization of the data that was collected to show all the diversity criteria.

Understanding the socioeconomic and demographic characteristics of an urban region is vital for policy-making, urban management, and urban planning. However, gathering socioeconomic and demographic data is difficult, often relying on costly and time consuming human-participant survey work. This project explored how the typefaces in public areas could be used to measure the socioeconomic characteristics of a neighborhood. Using a deep learning model trained to recognize type characteristics, this study analyzed 59,515 typefaces images from 748,471 Google Street View images to measure how type correlated with household income in each neighborhood. This work concluded that typeface can act as a metric to measure socio-economic characteristics.

Open-source fonts are ubiquitous throughout the Web, and yet we understand very little about the implicit connotations and underlying perceptions that users have of them. Nor do we understand how these preferences vary across cultural barriers. What’s Your Type seeks to understand these font connotations and perceptions through a web-based survey. The survey asks users to pick the font that best matches an emotive adjective—including “playful,” “informative,” and “trendy.” By gathering user responses, What’s Your Type crowdsources a database of font preferences, searchable by language and location of the respondent.

description

description