Summer of Math Exposition

Learned something new today. The explanation is clear, the pace is good.

Great animations and format. The content was well explained, and each portion of the video felt intentional. The progression was also rewading and a viewer is left with a full image of their query. Great work!

Fantastic. Interesting topic together with clear narration together with clarifying animations.

there is not much I could suggest for improvement other than that the ‘morph’ animation of text is more distracting than it is helpful. A simple cross fade would be better in the times you use morph. But not all the time. The fade animations on text at about the 28 minute mark is great and you shoudl have more of that kind.

One of the biggest questions I have with this video is why use a DAG, specifically acyclic, many variables are extremely cyclic, and are commonly called feedback loops, ignoring these seem like a huge oversight, especially if you can’t include them in the analysis. I understand it certainly makes things simpler, but there was no justification or reasoning for it. The examples covered seemed a little contrived, except for maybe the last one.

Overall good video, clear conscise and easy to follow. If a little simplistic, but that is probably not a bad thing.

this was really great and well put together visually and I found myself vibing with the music. it really set the stage for the explanations you gave. and the title: “A Painless Introduction” was spot on, in my opinion. however, there was quite a bit of terminology to learn if I was seeing this for the first time. maybe a glossary at the end or a review could be helpful.

the sequencing of ideas was good, where the ideas built on each other. then the breakdown into chapters was great. it gave me a good stopping point to reflect on the ideas that were presented and think about what might come next. the surprise of the “not so fast”s were well placed and useful. on the whole, it felt more interesting rather than a magical experience.

it would be cool to see more complicated and real world examples of DAGs used in more examples of research. by the end, I was wishing I knew more about how DAGs were constructed in the first place and some rules of thumb for choosing what goes into it. there were a lot of X, Y, Z, U, C, and M but it might have been nice to see more varied examples of what this might be in practice.

“Control for ” always sounds like it’s a deliberate action, but one thing that may be worth doing more than hinting at is that sometimes it happens unintentionally. With your two-disease case, you said casually “if we only look at people in the hospital”. But that is precisely an action that someone might literally take - if they do their survey by collecting stats from hospitalizations, they are unintentionally creating this control. Don’t know if that’s something worth spending a lot of time on, but it struck me as a potential omission.

(Small verbal typo at 34:32 - you say that Z cannot be confounded by Y, s/be U)

I thought this video was great, and about a concept I always wondered about (how to determine correlation vs causation). It was very in depth. Maybe this is just a personal preference but I would have loved a video that was shorter and highlighted the main important points including their implications. I think the manim animations helped a lot but maybe I would have liked some variation or other visuals besides manim? These are just ideas had while watching—anyway great video!