Template to run a 90 min intro to ML workshop

Poster for workshop. Find files on Github.

This March, my friend, Jihyun Kim and I decided that we wanted to share with students some of what we’ve learnt in the realms of data science and the Anaconda package. This workshop was a part of series of Interactive Media workshops at New York University Abu Dhabi.

After some researching (googling around) we couldn’t find any general python+ML introductory workshop material that was of the required length, accommodating all skill levels, and engaging. The ones that were online were either very plain and uninspiring or too complex for one 90 min session.

Given the limited time, we set our goal to create a workshop that explains the basic data science stack and inspires students to learn on their own. Our target audience was someone who understands the basics of programming. Through the workshop we hoped we could spark conversations about the ethical issues around automation, trusting black box algorithms to make decisions, and inherent biases in algorithms. Lastly, we also want to share our material with you to make it easier to teach ML.

All of the material is on Github free to use or contribute to.

We started off in the first half with explaining what Anaconda is and how to set it up. We then went over a quick review of Python and an introduction to some of the Pandas functions needed for the second part of the workshop. The setup instructions are in the README, coding starts in the part1 notebook.

In the second half, we worked through a fictional story: In the country of Arstotzka, things aren’t going well and the country’s food supplies are running short. If citizens want to eat, they have to send a letter to the Department of Food Supplies. The officials judge their requests and either reject them or send a slice of pizza.You’re the department’s software engineer and your boss told you automate request approval with this new thing called machine learning… Link to the complete notebook and to the skeleton notebook. Our dataset is the Kaggle Random Acts of Pizza dataset, which we modified for the workshops purposes.

The workshop went super smoothly, and it was awesome to see so many people, especially freshmen, who only had abstract ideas about machine learning. Now I get why teaching can be such a satisfactory work. We also learnt a lot that we’ll incorporate the next time we would teach a workshop.

Our general workshop advice:

We would love to hear if you found our template useful. Feel free to use it in any way!

Software Engineer in London • gaborcsapo.com

Software Engineer in London • gaborcsapo.com