In my (grad) classes and bootcamps, I show very little slides and spend most of the time running through exercises via live coding and hands-on exercises. For slides, I typically start the instruction by showing what is possible in order to get the students excited, and from there, discuss the topics that we will cover and why they are important.
For the hands on part, I live code using a commented script in front of the students, talk about how/why things work, and then ask the students to complete short exercises. Simply, 5-15 minutes of my instruction and overview depending on the topic, and then ask them to work through increasingly complex simple code challenges (3-5 questions), in which they get 10-20 minutes to complete on their own.
During the hands-on work, I am walking around the room answering questions as needed. If you wanted a variant of above, I would start out with live coding in front of the students (laptop down), run through some guided examples with the students, laptops up, and then have them complete a few code challenges on their own.
Regardless of the approach above, I repeat this process for the entire class meeting. It may seem repetitive, but I have had outstanding evals over the last two years with this approach, and one that has worked for both R and python instruction.
Of course, as the class meeting starts to near the end of the time allocated, I attempt to create a code challenge or two that pulls everything together to round out the core learning objectives and tie up the material that was covered.
I hope that helps, but from my experience, I would stay away from slides. I have tried it, and while it may work for others, I have found that approach to be disruptive.