Wednesday, March 27, 2013

wtf is logistic regression

i still read nate silver's blog from time to time, and today i noticed that he's doing an analysis of how the issue of marriage equality is trending in america.

i wonder about his methods, though.  i mean, literally.  i like the conclusions he comes to, but have trouble making much sense of how he gets to them.  even tried reading the wikipedia article on 'logistic regression' but still find myself wondering what the hell it means.

it got me thinking though, about whether all the mathematics he's talking about might be the sorts of calculations our brains are all capable of - to varying degrees, decidedly fuzzier degrees, but calculations we at least attempt to make, every day.  such as when we cross a street, when we think about the likelihood of a car coming, the probability that they'll stop for us, or if they've got a red light, the chance that a car from the intersecting street might come whipping around a low visibility corner.  we consider all sorts of factors, is it raining, is it approaching rush hour, is it a friday night, etc, etc.  Ultimately, we make the decision that we probably won't get run over, and hopefully, our calculations are accurate enough.

still, i wish i knew more about this math stuff.  it does seem quite a bit weightier than just saying, ok, here's what i think...

anyhow, he talks about how this country's trending towards liberal acceptance, and away from conservative bigotry.  or as i've been saying, society is moving forward.  the speed may vary, but doesn't it always move forward?  conservatives may hold it back some, we may stumble over their foolishness left and right, but do they ever turn society around?  regain any significant lost ground?  seems to me, they always just begrudging keep dragging along behind everyone else, eventually even moving forward themselves.

Neil Degrasse Tyson does this talk, though, about Baghdad circa 900-1200 AD.  About how much of modern science and mathematics are named in Arabic, due to so much of it being discovered in that part of the world, during that time.  and how it all came to a grinding halt, when there was a surge of religion in society.  A surge of conservatism, if you will.  It does seem that society very much got turned around there, at that time.  For one thing, the scientific discoveries flat out stopped.

It's difficult to imagine how anything like that could happen here though, at this point, short of maybe a zombie apocalypse.. but my knowledge of world history is a bit spotty (ok, it's crap) so I'm not sure how often society ever turns around like that.  I can't even think of any other examples, excluding wars, progress being stamped out by military conquest.

I think Tyson may be concerned that the same could happen here, but that's not what I'm seeing.  No idea if Nate Silver's math would back me up on this, but from what I can figure, religion is on the way out, in a huge way, taking a whole lot of its baggage with it.  This sort of thing takes a while, though.

That it's only taking a matter of decades is actually quite rapid for a change like this.  I guess science has been especially convincing, this past century.

Tuesday, March 19, 2013


Internet privacy is a fallacy.  This is a public place, but I think many people have trouble understanding the concept that while they're physically sitting in the privacy of their own homes, they are virtually putting themselves out there, amidst millions of others, like visiting a sort of global shopping mall.

This isn't to say you can't wear a disguise, or try to hide in virtual dark corners and alleyways, but just like doing so literally, it still isn't private.  A part of you is out there, in digital form, in the digital open world.  This means some of the expectations we have of privacy are a bit silly, and some of that paranoia about it being breached is a bit misguided.

This is not something anyone is doing to us, not the government, nor corporations - they may use the data trails we leave everywhere, but it's really not so different than the way they use real world data, in the ways we're more familiar with.  From which shopping sales we take advantage of, to the fingerprints we leave behind.  A lot of it is being used academically, as well.   It's all just out there, as a natural result of how the internet works, and it only makes sense that people try to make use of it.

Now, if there were any evidence that we were being watched too closely, too specifically, sure, that could be scary -  but much the same way it would be, if we found out they were tracking our physical world behavior too closely.  Being stalked, audited, stopped and frisked, etc.  We do have places like Facebook, that blur the lines, where they ambiguously give and take privacy, but it's like a restaurant, where diners don't seem to realize that they're patrons of a privately owned business.  Where one person's breach of privacy can be just the exposure another person wants out of being there.

I think maybe the problem lies more in the common understanding of what sort of place this is, and really, what sorts of places we go, all over the internet.  Which places are public, which are businesses, and where we draw the line between that and our presence there ..and I'm not sure some people even understand what the internet itself really is, and what it means, just to be logging into it.