darkxanthos

Wait what? Public opinion surveys (if that's what you meant) are absolutely still subject to error. This is especially due to getting a representative mix of all of the different populations.

Even without those issues, you still have the issue of people you survey not telling the whole truth... or surveys being poorly designed so that the question people think they are answering is not what was really being asked.

Data is hard. Just because you have some survey data and (maybe) did something statistical doesn't at all mean you should listen to it.

I was really surprised the article didn't mention this. To be fair it's apparently something that has been evolving for decades, so it wasn't so much prediction but they definitely aren't inventing something wild. https://ourworldindata.org/fertility/

PyMC is roughly equivalent to Stan. I think what you're getting at is more the applications of these statistical models. Stan can be integrated into Python via PyStan.

I think it's the most exciting. I love seeing nature effectively defending itself.

You're not a cynic, that's literally their business.

I seriously had no idea Heroku was still even actively being worked on. Nice to see a much needed upgrade!

Thanks for commenting! The whole technique is foundational Bayes. Doesn't use MCMC, but still bayes even with the use of credibility intervals. I don't make much of a use of priors to your point and that can definitely be seen as somewhat controversial, but each window is aggregating information on all of the hypotheses given all of the data. That's why I qualify it as "bayesian". How do you typically qualify bayesian vs not?

To your point of the issue with false positives... yep! :) It's a very simplistic naive approach. I read over the article you linked to and I think the results are fairly similar except that my (again very simplistic) technique finds many change points if they exist... and possibly even none. So what I'm doing generalizes more.

The sequential nature of what I'm doing bothers me too, so I'd love pointers to other articles doing something similar.

+1 to this. I recently just started toying with the idea of going sugar free and ketogenic... so I started trying foods without many carbs but wasn't even committed to it. All of a sudden I noticed I had lost 5-7 pounds in 3-4 weeks. So I got excited and kept it up... I'm not losing weight at the same rate but now about 2-3 months in and I'm down about 25 pounds. It's ridiculous.

I haven't noticed the energy or clarity improvements in myself and my wife hasn't seen that in me either. That may be due to sleep as I'm going to school full-time, training for the Chicago Marathon, and working full-time. Given all that maybe it's just great that I can function. :)

Either way! The changes have been amazing and dramatic. My biggest problem now is that I need to buy a whole new wardrobe but I can't yet because I think I still have another 15 pounds I'll blow through by Christmas. I finally feel in control of my weight and it's the best.

I think Lean Startup & Customer Development transformed the way I looked at everything I did as a software engineer. It reignited my inner-entrepreneur as well. Now I'm a data scientist and I work in Finance and I help get others in my company to have that same epiphany.

Exactly. I'm guessing most didn't fully read the article. What I got out of it was that the NSA was really concerned about being sure they were correct that Snowden didn't in fact officially raise any concerns and that all they could find that was even close to this was a single email about a trick question on a test.

One could argue the NSA could be trying to make sure there was no proof rather and that's fine. This article though only finds evidence of a single email. I'd love to see more but this isn't it.

How does bit shifting prevent typos as opposed to just explicit numbering?

A better example is the Monty Hall problem. Very few people find this intuitive.

A co-worker pointed out to me the other day that PyCharm has SQL querying and refactoring built in as well. Starting to use it as a data grip like IDE as well.

My professional advice to people who make these unmeasurable changes is: Then don't do them. The worst part about unmeasurable changes is you can't verify if it has a negative effect either and you're essentially saying "I'm focusing on making this change when it will have negligible impact on the business."

You have better things to do.

Salary.com, glassdoor and other resources are not random but they are useful.

One possible solution is that it may mean more sponsored content. Even in mainstream media a surprising amount of content is sponsored.

If you aren't paying for your content, your content needs to pay for itself.

Standard deviations assume normal distributions... but ultimately when you look for some two standard deviations from the mean, you're just look for an event that only occurs 95% of the time. A technique that works regardless of distribution is percentiles. Postgesql supports these now as well.

That's a really good point. I got so fixated on the jackpot but there's more to winning than 1st place. :)

In a world where hard core gamers will pay >$500 for a graphics card, this price will work fine for a 1st gen. It'll be a smaller audience at first but that will also help them out as they need to focus on their customer support for issues they find in the wild.

Anything that can help make this subject more approachable is alright in my book. In my experience, the hardest part of Abstract Algebra isn't the abstract part of it. It's the lack of examples and clarity around how different named concepts behave. The proofs aren't terrible even except for the fact that they require a familiarity with number theory that many (at least I myself) don't have. I love the subject though. Really interesting.

I don't think anyone has said this yet: Use your lack of time as a motivator to distill your side project down to the simplest thing that could possibly work (MVP). This is your opportunity to get really good at that. If you can't test out an idea in a weekend, you need more practice at this not more time.

With automated drones and Amazon brick and mortar stores the idea of large warehouses could be challenged. They could shift packages in an automated fashion to areas with more probable demand. Drones are crazy cheap as well. I can imagine hundreds of these in the sky at once in Chicago if not thousands.

All a UUID is really is an incredibly long random number (at least v4). I'm not sure exactly what you're advocating here.

Yup! I came here to post exactly this. Like you said, mathematically it isn't perfect but in reality it's basically rock solid.

If space is a concern, this time stamp could also be converted into ascii or the like.

This. It is so important to separate the practices associated with Agile apart from Agile. One practice I use a ton is retrospectives and really it's useful just as a way of codifying how much and how to experiment with and change team process.

It's common to refer to speed ups in powers of ten as orders of magnitude. Maybe that's it?

I think that this is flawed in the sense that most people (including the article's author) might misinterpret it. The paper assumes that we know when a streak ends. So given that we have had a streak of tails and that it's broken what's the probability of heads? That will be biased toward heads. I've only skimmed the paper, but maybe they ultimately mean this is the source of the incorrect bias in the layman's intuition?

This exactly why string theory comes under fire. "Not even wrong" because that would imply falsifiability.

Probabilistic Graphical Models is a pretty good blend. Not sure how theoretical you want.

I took a step back and read Think Bayes by Downey and watched some of his youtube videos. Then Introduction to Bayesian Statistics by Bolstad is great once you're reading to deal with probabilities. Now I'm reading

* Building probabilistic graphical models with Python (Karkera)

* Mastering probabilistic graphical models using Python (Ankan)

* Probabilistic Graphical Models Principles and techniques (Koller)

Having skimmed (just started reading them) I'm very excited to continue to delve into them.