What I learned in 2017

Here are some things I learned in 2017, a summary of what I worked on the last year, and what I hope to accomplish in 2018.

A lot of them are things that have been said before. I think they bear repeating partially thanks to the lucky 10,000 effect. Another reason is what's obvious to you isn't necessarily so to others - the most heavily-trafficked page on the site by far was Installing OpenCV in a virtualenv, hardly groundbreaking research.

I include my future self in this: the page I visited the most in 2017 was generating Django secret keys. I wrote it, and I still refer to it all the time. You in the future is a different person.


Consulting

2017 was a big year for drones. I consulted primarily on the data processing side. I got to work on some really fun projects.

Many of the interesting things I helped build were photo-based: panoramas, orthomosaics (fancy name for maps), image recognition. Others used photos for more complex products: photogrammetry, point clouds, 3d modeling, video processing, etc.

A big highlight was building out an automated aerial mapping data pipeline. I worked with a client collecting drone photos and video from sites across the country. We were able to consolidate their process and automate much of it. We got a ~5x improvement in throughput over manually processing the data.


Speaking

I gave my first conference talk at PyOhio and reprised the talk at Milwaukee Code Camp. It was a blast.

I also gave my first, second, and third presentations at a meetup. These were fun too, and I made some friends working on really interesting stuff as well.


Personal projects

I created DIL, software for processing drone imagery.

I also created Drone Stable. I started showing it to pilots in October, and I've gotten a really good response. Here's what it does:

Visit dronestable.com to check it out.


Goals for 2018

  • I have not made as much progress on the computer vision cheat sheet as I would like. I want to write at least one item per month for it, either adding an example to the list or creating new items.

  • Continue consulting with drone companies. I think businesses can get a lot of value out of machine learning and automation. If you're one of those people and want to talk, email me!

  • Keep up on public speaking - I would like to give a talk at two conferences this year.

  • I am thrilled about the feedback I've gotten on Drone Stable. I want to dedicate the time it needs to really grow in 2018. It will likely be my biggest project for a while. There's certainly enough shaky footage around. I hope it can start to pay the rent :)

  • Publish the 2017 year in review on time.

See you in 2018!