caipre

It doesn't seem that implausible for a program to be reading through a dataset and issue a notice for a measurement above some value. If the dataset encoded years in just two digits, it (almost) makes sense that the program would report it 100 years off.

Imagine:

    with open(file) as f:
        for record in f:
            year, month, day, time, magnitude = record.split()
            
            if magnitude > 5:
                raise_alarm()

https://pocketearth.com/

In case I just missed them, are there some comparable projects to this? Is there an established library in the space?

Published today, and showed up at the top of my NYTimes Now app

Steve, thanks for all your work on this and with Rust. Your attitude toward teaching and your writing ability go a long way in encouraging people to learn and in having the lesson be worthwhile.

    > Lidar data is collected by low-, slow-flying aircraft with equipment
    > that shoots millions of laser points to the ground. ... It is
    > possible to strip buildings and vegetation from the images, so that
    > only the ground is shown. In the Willamette River poster, the shades
    > of white and blue show elevations. The purest white color is the
    > baseline, (the zero point, at the lowest point near Independence on
    > the upper part of the image). The darkest blue is 50 feet (or higher)
    > than the baseline.

Edit: The original submission (see dang's post) references a similar set of maps[0] for the Mississippi River, which are equally beautiful and even more impressive.

[0]: http://www.radicalcartography.net/index.html?fisk

* Why is constant time desirable here, if the operation can be quicker for some inputs?