This weekend South Africa had its very first Raspberry Jam meetup that featured a video call with Raspberry Pi founder Eben Upton.

We were not only there but one of the lucky few to ask Upton a question during the brief Q&A session, so we asked him what his all-time favourite Pi project is after seeing thousands of them over the years.

His answer is the machine-learning, cucumber-sorting machine created by a farmer’s son in Japan.

You can read about the project – which also uses TensorFlow – on the Google Cloud Platform blog, or you can listen to Upton explain it below.

We’ve timestamped the video to the beginning of the conversation about the sorting machine, but you can also fast forward to 16:57 to hear the full answer which includes a “serious” project too. Upton classifies the cucumber sorter as “whimsical”.

If you did fast forward to listen to the full answer, you got to learn about the practice of ballooning, in which Pis and cameras are strapped to weather balloons and sent up into the atmosphere.

BinarySpace, a local makerspace, did something similar in 2015 by sending a balloon 32 kilometres up as part of the Global Space Balloon Challenge.

On a smaller scale Upton points out how the Raspberry Pi Foundation is training teachers to create these balloons as a class lesson, which was first seen in secondary schools and is now even being used in secondary schools.

But back to cucumber-sorting and you can see the machine in action here. While it may seem a bit on the slow side, it saved the farmer up to eight hours of manual labour doing the process by hand.

[Image – CC 2.0 Håvar og Solveig]