Lenny Bogdonoff, a New York-based software engineer and graffiti artist, is creating the world’s first digital genealogy of street art. He played and won a Pioneer tournament while developing a set of machine learning tools for his project, Public Art.

By gathering photos from around the internet and using machine learning models to identify street art, Public Art aims to digitally preserve murals around the world.  

Where did the idea for your project begin?
A few years ago, I relocated to China. The move was officially part of a study abroad program, but I really wanted to go to paint graffiti. I had a connection to a local Shanghai artist with an open spot in his street art crew and by my third night in the city we were rolling out and painting street murals — it was such an awesome experience.  

When I finally left China, I was grappling with some big questions about street art, its role in society, and how I could help it achieve greater public recognition as a serious art form.

That was the beginning of the idea — what if I could create a genealogy of street art so, like other art forms, it was searchable and citable for the first time?

What was the next step?
I majored in the humanities at college, so I didn't know how to code well. I decided to invest my time learning to build small apps and experiment with how that could help me answer this question.

I settled on two major areas that I needed to tackle, in creating the genealogy:

   • Creating a way to technically analyse images
   • Crafting and managing a digital archive

I’m currently focused on surfacing images that are geotagged with a longitude and latitude point and scraping Instagram. Eventually, through machine learning tools, Public Art will analyse images and start to aggregate data about the forms that street art can take.

What does that mean in practice?
Every day, people take hundreds of thousands of photos of street art around the world. If these are logged in Public Art, important details like when the artwork appeared, when it was erased, and what neighborhood it was painted and photographed in, can be logged. The images can also be compared for stylistic trends and influences. Having these digitally archived creates a central location where images can be viewed.

Why are you focusing on machine learning?Analysing art is a passion project, but the skill set I’m developing in machine learning is more tangible for changing the world. These technologies are valuable, and whatever I develop for Public Art can be implemented by other industries, too.

Why did you decide to play Pioneer?
I heard about Pioneer online and decided to play the tournament as a method of experimentation. Having a level of social validation is really important for creativity, motivation, and energy, and I wanted to test out ideas with a like-minded community. Just knowing that others found my project interesting was a nudge to keep exploring. I’ve stayed connected with the Pioneer cohort through Twitter and Slack.

What’s next for your project?
I’ve just launched Public Art. We currently have photos from 180 countries (and counting) in our digital street art museum.

Since I have started the scraping, I have gathered photos from all over the world, including places like Azerbaijan, Cambodia, Nigeria, and Kazakhstan. My current workflow for getting images, results in approximately 10,000 new geo-tagged photos a day, so I'm working on some improved ways of browsing the content.

I’ve also developed training models for classifying street art so that others can start applying the principles. As a whole, Public Art is still in the experimentation phase and I’m currently exploring different ways that the tools could be applied across commercial art, education, and the not for profit sector.  

What are some other crazy projects you’ve worked on?
For fun, I made a tool to finish long New Yorker articles, co-organized hackathons for journalists at major media companies, am an avid follower of graffiti and painter, and once I even co-created a break up story hotline (check it out here).

Top tip for playing Pioneer?
Don’t necessarily play to win. There is value in getting feedback on the areas of your project that won’t necessarily be the quickest to progress.Think about the big picture of how connecting to the community will help you reach your ultimate goals.