I mind less non zero axis graphs than no unit graphs.

Apple is doing this for ages....Like how they demonstrate Swift is faster than Python, lol

Starting the axis somewhere other than zero may be forgivable for some types of graph, but bar charts aren't one of them.

For a keynote, I'm ok with the no axis graphs. It's technically not even trying to be a "graph". Just a comparison graphic. But units are a must.

In this case the information is more misleading because it is on a graph than it would be with just numbers. How can it be a good "comparison graphic" when the graphic part of it impedes my ability to make a comparison?

And the first image doesn't even have units…

>What's with the non zero axis graphs?

It's how you emphasize a change in the series, and is an acceptable method provided you aren't intentionally misleading. Context matters.

If you look at the presentations, they are comparing only 2 or 3 values, so graph itself is meaningless because visual delta between two values does not mean anything.

How so ? One is showing power consumptions in milliamperes, the other is score marks in GFXBench - both of which should be quite meaningful.

Ofcourse you have to know what GFXBench is measuring.

What information is the graph actually showing? That number X increased, sure, but not in any way by how much. By not having a labeled Y-axis, and not having the y-axis start at zero it is basically meaningless from a communication standpoint. It might be effective from a marketing standpoint, but only in the misleading way.

They made no effort to be truthful. Now I don't believe anything in that presentation.

Printing the actual data values on the bars seems to contradict "no effort to be truthful".

Just accept the bar graphs as the decorative ornaments that they are in the presentation. Even stripped of scales or origins the three bar graph manages to convey relative improvements, very much a Monty Python Black Knight move on its part.