willis77

Let's call the whole thing off

We (Kaggle) run a data science jobs board (https://www.kaggle.com/jobs) that gets a few data engineer listings from time to time. Not all of these are active, but you may find a few interested companies via - https://www.google.com/#q=site:https://www.kaggle.com/jobs+%...

Upvoted from my sensibly balanced Craftsman home while eating crusty artisanal bread and drinking pour-over Blue Bottle coffee before I head to work in my 2012 Volvo XC90 while listening to NPR.

Kaggle has a free R/Python/Julia/Jupyter environment with lots of code to browse. See:

https://www.kaggle.com/kernels https://www.kaggle.com/datasets

Some bikes that are hard to ride no-handed have "notched"/"indexed" headsets that are to blame. Small indentations in the headset prevent minor corrections from actually turning the wheel, making it hard to let go of the bars.

http://www.sheldonbrown.com/headsets.html#troubleshooting

(see section on "Indexed Steering")

"Sit down, little Timmy. Grandpa wants to tell you a tale, a tale of yogurt parfait in tiny cups as far as the eye can see, a tale of padded deluxe banquet chairs and Starbucks coffee pods from worn hotel percolators..."

Sounds glorious. In the U.S., figuring out taxes is like:

Take line 39b, multiply it by your age as determined by worksheet 6689. If you are between 41 and 43 years of age, enter '0'. If you were not subject to the Railroad Act of 1938's exempt dependent pension clause (this is not common), add 13 to this number in hexadecimal, else, subtract 2 and multiply by 1.388688. If you did not have healthcare for more than 14 minutes during the prior year and line 17 is 4 or greater and you live in a zone designated as "clown free" by your local municipality, draw a unicode snowman on this line and prove Fermat's theorem in the margins of Schedule Q...

I'd like to see the mathematicians ascribe an expected value on the bitter, icy glare of your envious colleagues as you eat that sweet victory lunch.

It was a joke that nobody found funny.

https://www.youtube.com/watch?v=eSxlfLTZu0k

Cornell? It's pronounced "colonel" and it's the highest rank in the military.

The fact that they haven't implemented a useful ratings sort drives me nuts.

http://www.evanmiller.org/how-not-to-sort-by-average-rating....

"Oh, him? That's Chuck. He just runs our commits through a minifier and takes credit for the lost lines. Means well, but he's kind of a jerk."

"A data scientist is a person who is worse at statistics than a statistician and worse at software engineering than a software engineer."

- Me (Data Scientist)

In the US our office of tax simplification would also need an oversight committee on tax simplification regulation, an office of tax simplification administrative affairs, a center for tax simplification reform, and an office of tax simplification simplification.

"It looks like you're trying to write a potato. Would you like help?"

Looks like they forgot to test for negative values. I put in -200.00 and it seems to have withdrawn the money from your account without problems ;)

If you haven't seen it, Fibonacci Flim Flam is a fun read - https://www.lhup.edu/~dsimanek/pseudo/fibonacc.htm

This problem is tightly coupled with issues of correlation vs causation. In real data, correlation is mostly transitive (there are toy exceptions - https://terrytao.wordpress.com/2014/06/05/when-is-correlatio..., but they are just that). This means that if you want to predict something, and that something is correlated with some some uncomfortable association, trying to predict the something without the uncomfortable association can leave little residual behind.

For example, if hair length is societally taboo for gender prediction and I make an algorithm that uses a "politically correct" determination using XY chromosomes, I have also made an algorithm that correlates with hair length. Moreover, if I try to statistically correct my algorithm so that it does not correlate with hair length, I end up with an algorithm that works on the tiny leftover residual created by people who buck the trend, i.e. one that's much more likely to be wrong.

Algorithms find both correlational and causal factors. If 9/10 men are from Mars, and you tell me you're from Mars, it is often via correlation that the algorithm labels you a man. You are not allowed jump to the assumption that, say, the drinking water on Mars is turning people into Men.

This is one of those Calc II problems where it turns out the profit limit as the number of frontrunning funds -> ∞ equals something like e/log(2)

If you enjoyed this, don't miss Moritz Hardt's entertaining blog post on Wacky Boosting (linked from the forum thread):

http://blog.mrtz.org/2015/03/09/competition.html

Hey man, he took those MIT engineers to school. Despite decades of focused dedication to the study of robots, control theory, gait and dynamical systems, they probably never considered just making their robot work more better.

Cycling is great for giving an appreciation for what 20-30mph really feels like, without the car disconnect. Same as your sprinting example: go get a bike, find the steepest hill you know of, pedal as fast as you can down the hill (this is going to be 30-40 mph), then ask yourself if you'd like to hit a brick wall in that moment. Most people who are otherwise comfortable riding without seatbelts would be peeing pants and grabbing brakes long before 40mph.

It only shows when the trip/meter starts.

As a data scientist, I draw inspiration from the infinite depths of understanding. Life is merely the unfolding of a series of recursive recommendation engines, each one AB testing the local gradients of human emotion. You think you are just buying a pair of socks online. Foolish mortal. That transaction was a multi-armed map-reduce set into motion long before the internet existed. Data science is the internet of things, it is big data at the speed of entropy, quantum mechanics at the scale of desire.

In data we trust. All others bring EVEN. MORE. DATA.

He probably fedex overnights them on 32lb bond California Redwood sourced stationary.

> People often ask how my business is going by asking if it keeps me "busy". They're using the wrong metrics.

Or just being polite? Or following norms about not asking about money? Or just using "keeping busy" as a construct for "not unemployed"?

> a measured amount of properly ground coffee.

And it's not even doing this well. More accurate to call it "a measured amount of ground coffee, 'aged' 8 months in a warehouse."

The problem with ascribing dollar values to data (a.k.a. facts) is that you can make up a million untestable hypotheses about how data could theoretically have worth. To give an extreme but illustrative example: let's say that by typing this comment to you, I raise the temperature in my router by a millionth of a degree, which alters the wind pattern in my room, which causes butterfly to flap its wings, which... you get the point.. causes a hurricane in some remote part of the world. Does this make the "expected value" of my comment negative billions of dollars?

You can back this argument out to less extreme versions and see that it's very difficult to economically value a fact. E.g. what if the traders named in this article had 100X the capital on the line for the same trade. Is the data now 100X more valuable because of the added economic value they derived from it? What if their trade netted them more than the change in market cap?

Ahh, but he only reads the first eight letters, so all he got from your email was "Greeting"

Her blog is also a great resource to pick up some of the more useful parts of the language.