HackerWhat becomes visible when you change representation?
An average, a spectrum and a geometric diagram each foreground a different structure. Before asking which view is intuitive, ask which comparison you need it to support.
Questions in this reading · 6
What disappears inside a national average?
Countries can improve on average while wealth remains concentrated within one country. Compare the unit of observation before comparing the trends. Wealth shares and life expectancy measure different things.
Compare original passages and timestamps · 2
1% of America has 40% of all the nation's wealthInequalities within countries these bubbles show country averages, but I can split themWhat does a visualization let you notice?
A color map controls how differences reach the eye. Moving bubbles can expose patterns over time. In both cases the encoding influences what viewers can detect.
Compare original passages and timestamps · 2
an interface between your data and your brainhaving the data is not enoughCan changing representation reveal a solution?
A different representation makes certain relationships directly visible. Compare which calculation each representation makes easier and which information it preserves.
Compare original passages and timestamps · 2
choose the appropriate
data structure before you start tinkering
around with the algorithm.that right-hand
side that I'm trying to get is actually one
of these columns.Is touching every pixel the same as filling the plane?
A sufficiently fine Hilbert approximation visits every grid cell on a screen. Magnify a cell and there is still empty space beside the line. The square-filling curve is the limit of the sequence, not its most impressive screenshot. A talk on infinite sums offers a second check on intuition: visiting the neighborhood of one is not enough. For any tolerance, the partial sums must eventually stay inside it. These are different limiting objects, but both replace “looks more and more like it” with a condition one can test.
Passages and timestamps · 2
The output of the Hilbert curve function evaluated on this input is just defined to be the limit of those points.no matter how small the distance you choose, at some point down the list, all the numbers start falling within that distance of x.Counting boxes along a coastline asks an empirical question: how does the occupied count change with grid size? A log-log slope estimated over one scale range can change over another. One nearly straight segment does not establish the same scaling law down to arbitrarily fine detail. When a picture seems conclusive, ask what it actually shows: one finite sample, a comparison of scales, or a claim about a limit?
Passages and timestamps · 1
this dimension, at least as I've described it so far using the box counting method, can sometimes change based on how far zoomed in you are.Which frequencies are present, and how was the signal recorded?
A Fourier transform describes a signal by frequency. Sampling and quantization describe how a signal is measured and encoded. They answer different questions.
Compare original passages and timestamps · 2
It pulls out the original frequencies from their jumbled up sumsthe sampling step is perfect it loses nothingWhere does an error in precision matter?
An approximation has to be judged at the output where it will be used. Compare how the two demonstrations distinguish useful accuracy from unnecessary precision.
Compare original passages and timestamps · 2
neural networks don't need the precision of digital computers.dither is specially constructed noise that substitutes for the noise produced by simple quantizationCan you reconstruct the area formula?
A formula is easier to transfer when a learner can rebuild the reasoning that produced it. Compare a memorized answer with a representation that makes the derivation visible.
Compare original passages and timestamps · 2
build mental models that you can use in other contextsthe sum, thought of as the aggregate area of those rectangles, approaches the area under the graph.Try it on your own example
Represent the same data twice. Which question can each view answer, and which question can it hide?