Source: Cam Peters • The Spectator

Ugly Produce: Coming soon to your doorstep

Published in
6 min readNov 9, 2018


In this article we describe how Topos’ transformative understanding of location is powering the strategic geographic expansion of one of our favorite customers: Imperfect Produce, a company that is fighting food waste by finding a home for ‘ugly’ produce.

Ugly on trend

Right now, ugly is cool. It’s cool in fashion (ugly sweaters, ugly shoes), on Netflix (David Chang’s foodie porn show “Ugly Delicious”), and even in produce. You may be scratching your head and asking “ugly in produce?” But, there’s a growing movement to embrace the ugly in order to address the issue of food waste in our country. Just because they’re ugly, 20 billion pounds of fruits and vegetables are wasted on farms each year.

Ugly produce stats

In the United States, up to 20 percent of fruits and vegetables are wasted because they look a little different, and grocery stores won’t accept them, whether due to size (too big or too small), color, or blemishes, even though these imperfections don’t affect flavor or nutrition. Take a good scroll through the Instagram account uglyfruitandveg to get a sense of what ugly produce looks like.

Here waste spans multiple dimensions: from the number of Americans experiencing food insecurity to the resources required to grow produce that will go to waste.

In the last few years, there has been a growing awareness of this problem and a movement to reduce waste by bringing ugly fruits and vegetables into the homes of consumers who are more than happy to snack on heart-shaped apples and cook with two-headed potatoes. Since 2015, the startup Imperfect Produce has been expanding their footprint in the U.S. and bringing ugly fruits and veggies at a steep discount to its customers.

Ugly produce for the win

There’s no shortage of reasons why the work that Imperfect Produce is doing is important, but in case you need a few more…

  • Cosmetically imperfect fruit may be more nutritious than it’s cosmetically superior peers, perhaps because plants rely on their antioxidant defenses to fend off environmental stressors that cause imperfection. (source here)
  • The company has also donated over 1 million pounds of ugly produce to food banks and donation partners–not only are they reducing food waste, they’re helping to reduce hunger in our country.
Topos — Similarities View

What is Topos?

We formed Topos to advance the understanding of location through the interconnected lenses of data and artificial intelligence. While there are well-known tools such as the United States Census that use manual techniques to collect information about different locations, using data and AI enables a dynamic, highly granular, and globally scalable understanding of place — an understanding we think is valuable given the rapidly evolving nature of cities and neighborhoods around the world. Powered by hundreds of structured and unstructured data sources, and pulling from a wide variety of technologies and disciplines — computer vision, natural language processing, network science, machine learning, statistics, topology, urbanism, data visualization and information design — we are creating a transformative platform for understanding urban life and the culture of neighborhoods.

The beautiful expansion of ugly produce

Imperfect Produce came to us in 2017 with a question: after a successful launch in the Bay Area and Los Angeles, they wanted to understand how to plan their next wave of growth by predicting customer penetration in neighborhoods across the Portland and Seattle metro regions before entering those markets.

By combining Imperfect’s sales data with Topos’s proprietary algorithms and location features, we trained machine learning models to predict the number of active customers, total customers, active customer penetration, and total customer penetration for every zip code in Portland and Seattle.

Revisiting these predictions six months after launch, we found the results to be surprisingly accurate. As the prediction was intended to be used directionally, we decided to measure performance by looking at the Pearson correlation between actual and predicted values. For example we found correlations of .77 and .79 (p<<.001) between predicted and active customer penetration for Portland and Seattle respectively.

In light of these positive results, this year, we expanded our prediction work with Imperfect to cover 18 additional MSAs (metropolitan statistical areas) that were candidates for expansion. Using the same approach with updated sales data drawn from Imperfect’s enlarged customer base, we produced predictions for active and total customers per region. (Predictions for active customers can be found below; we’ve changed the names of the metro regions and normalized the predicted values for this post)

We produced multiple predictions for customer metrics (average, per capita, per sq mile, etc) because assessing the potential of prospective markets requires going beyond total active customers, to understand factors that influence warehouse location, distribution logistics and marketing activities.

Imperfect is using our predictions to support their national expansion strategy to bring imperfect produce to more customers in the U.S. Additionally, Topos’ predictions helped Imperfect more efficiently plan their delivery routes (to save carbon emissions) and better understand customer culture–a key component for building their marketing strategy.

Enabling Scale

Topos is proud to have played a role in bringing Imperfect Produce to more regions across the country. By understanding location as a set of relationships rather than solely as a set of isolated points or regions to which metrics are ascribed, we are developing technology that enables companies like Imperfect Produce to strategically scale to new geographic regions. Stay tuned for more from us on how we’re using our AI to revolutionize the understanding of location, but for now, more “ugly” produce from Imperfect.


Disclosure: Imperfect Produce has been a customer of Topos since 2017.

This post is part of an ongoing series capturing different insights we generate while developing our platform. We would love to hear your feedback. If you enjoyed this article please share and 👏 a few times so other people can see it too.

To learn more about how we are transforming location intelligence, please contact us:




Transforming the way we understand cities with Artificial Intelligence | @topos_ai