All relationships are linear! - Shitty financial and science journalism

My comment is about data usefulness and applicability, not about visualization.

But if they are making up the same percentage of overall buildings because the amount of buildings is rising too that would be a lot different than if the increase is on a stable denominator.

Useless without knowing if post office counts stayed stable or up and down over time.

What movie has a female plumber?

Just because you don't understand other numbering systems doesn't make them fake.

It's very long, which might work against rewatching and finishing it, reducing overall hours

Ok what does that have to do with the graph?

Nope but Fox News is. So I guess they don't matter.

An absolute free speech policy does.

So no rules against libel, slander, direct incitement to violence, indirect orders to kill a la organized crime, direct commands to commit violence or crime?

Yea, foreign tourists are going to get on a plane to St Augustin, Tuscon, and Asheville from NYC for a trip, rather than seeing one of the three biggest waterfalls in the world on a bus ride. It's a 7 hour bus ride that leaves from Midtown in the morning, does lunch at the border, a quick boat ride, a walk around, and then a bus ride back.

They're just gonna be like, "hey I'm doing a US trip to NYC and SF, let me just jam in a hop over to Montreal including border control and a quick pop over to Cancun (where flights for in demand times are 1200 dollars or more) and then a bus to Merida with two kids with a single entry visa. I'll also time a red-eye flight to New Orleans for super cheap, because my once in a lifetime trip to NYC is PERFECTLY timed to check out a cheap flight to NOLA, and I'm sure there won't be any flight delays for my connecting flight back to Korea that would be at my cost if I missed it."

Mostly they just do NYC with day trips to Niagara Falls and DC, and then SF with a day trip to Napa maybe, or LA to SF seeing Big Sur. MAYBE they do Chicago, but really unlikely, and if they have kids maybe Disney. Nobody is going to second tier cities or tiny college towns like Portland, Nashville, Memphis, Asheville unless they have friends there or they are like huge Elvis or Dollywood type fans.

People are FAR more likely to do nature trips as well then some small city...Niagra Falls, Grand Canyon, Yellowstone, Yosemite. It's like you only travel to cities and stay in hostels and aren't seeing what people actually do.

There are tons of buses that go to Niagara Falls for one day tours. And it's the same if traffic didn't exist maybe.

More importantly, Niagara Falls is a top 3 in the world sight for waterfalls. Noone's trying to visit Niagara Falls the city...

There aren't even ten real cities in the US worth visiting for an international tourist...

How many an easy bus ride from NYC or Toronto?

I think for many consumer products, like hair care or tampons, this is true. It becomes insidious when it's real estate, education, accommodations, etc if unchecked.

This is why we have laws around this. Let me guess, you think markets correct all inequities?

Absolutely not and I think it's fairly obvious it wouldn't. This was tried for education and housing and because of historical inequity and cultural in group bias of systems for a majority it doesn't work.

Even workplace or academic institutions that just have policies that appeal to white majorities can enforce that. It's trivial, but even not having say, vegetarian or halal items can be a blocker, and it's "race blind" to be fine not having it.

Clearly there's no strict line. Just like a white passing black person crossing the color line in Jim Crow, racist systems aren't absolute.

I'd say if there's a vastly disproportionate discrepancy it's worth checking. And I'd say if it's around things like housing, or education (rather than say, hair care items) it's more salient.

Did I say directly proportional? Stop strawmanning my argument.

Is the concept of machine learning making a racist assumption and enforcing racism alien to you? It's pretty widely discussed.

If companies ever backchecked their algos for mistakes or systematic bias, I might not be against it.

Ok so generalizing from a small sample size and then using race or race proxy demographics. Do y'all seriously not see the issue?

Racism is ok if you find a proxy. Got it.

This is like when politicians carve up districts based on other factors to proxy race. The model is definitionally racist if it continues to fuel racial segregation.

If the algo doesn't advertise to non white people, how would we know the problem is engagement. I'm trying to lead y'all through a line of logic that ends with the idea that outsourcing racist activities to an algo isn't not racist.