Catching Attention

An archival visualization that traces the Smithsonian’s digital collection of postage stamps from their introduction in revolutionary America to where they begin to live more permanently in modern day postal system and culture.

Create a visualization tool to make the vastness of the Smithsonian Institution’s Revolutionary-era data more digestible, pulling out narratives and highlighting stories hidden within the Open Access archives.

In investigating groupings of data within the Smithsonian search API and various Revolutionary era-specific Hugging Face datasets provided by the institution, I found stamps dating back to 1765 that prompted historical research on the significance of the Stamp Act in American history. After discovering its magnitude marking stamps as a key symbol of resistance and tension leading up to the American Revolution, I became interested in visualizing the National Postal Museum's vast collection of stamps within an online interface. Stamps in modern American society are an interesting example of something used for both practical purpose and personal collection - I wanted to explore how their introduction as such a point of tension in American history may have influenced their journey to this place over the hundreds of years since then.

I collected “stamps” and “covers” object type data and metadata from Smithsonian API, and used their titles to pull out themes/theme buckets and categorize stamps by decade. My initial visualization used a timeline and bar graph to explore the frequency count of stamps that fall into these themes throughout decades. I gathered historical research to supplement what the theme frequency counts were revealing about each decade and contextualize these trends within early American history.

In exploring where stamp images can fit into the website and how they might be displayed effectively, I used machine learning models like SigLip and OWLv2 to process the collection and standardize the images displayed. This enabled sorting images by similarity and cropping to the artifact, standardizing the visual information displayed and pulling out sub-clusters within each already grouped theme. I also decided to extract more metadata to see trends in colors and materials over time.

With this final iteration, I integrated my timeline with scrolling to allow users to easily travel through time while exploring the collection. I then rewrote my historical context to expose more word trends over time, and created filter buttons using words and colors to enable discovery and allow users to drill down to make sense of smaller historical trends that might get lost in the overarching views. I also created an individual stamp view for users to zoom in on a single artifact and view its data isolated from the whole.

America's Stamp Collection

Case Study Coming Soon