nnm

There are two facts A (my tool can help people use FB less) and B (FB banned me), and then tell people A caused B, i.e. A --> B. In reality, it could be C --> B and C (I collect user data in a way not allowed per FB ToS) is not mentioned at all in the story-telling.

Can't believe a comment that openly discriminated China-born US citizens is now a top comment. The whole logic behind the argument (around China-born US citizens) is so weak.

It's interesting to observe that whenever R is discussed, someone talked that python is better than R in this or that area.

Another reason why bitcoin should be banned.

"At their core, data scientists exist to create business value with data."

For some DS, this is true. For other DS, they are there to create or maintain a rigorous process so that we can reliable causal inference.

Example: how effective / safe is the covid-19 vaccine on trial. Thinking most of the statisticians in FDA and pharmaceutical companies.

Big money spent by bureaucracy does not have a reputation of not being wasteful.

A sensible capital gains tax should remove the inflation part.

Say you buy a stock for $1000 in year 1. Suppose monthly cost for living in that year is $1000.

In year 10, you sell the stock for $2500. Suppose monthly cost for living in year 10 is $2000. Your investment essentially just beat the inflation by $500 at year 10. In other words, real capital gain is not $1500.

The total crash rate in Tesla cars is not necessary less than that of say Prius cars.

Comparing Tesla cars crash rate with that of the overall population is dishonest:

1. drivers are biased population 2. the age of the car is biased

The Tesla Safety Report is so misleading:

1. The accident rate does not take into account of drivers age, credit score and prior safety record, or the age / value of the car.

2. Most people only turn on autopilot when driving is easy (e.g. on a highway).

Every time I was forced to pandas I miss R data.frame and data.table.

a few books by Hadley are really the best. For folks on HN, the "Advanced R" [0] by Hadley is actually a good start pont

[0]: https://adv-r.hadley.nz/

The devil is in the details. I am not sure it is an apple to apple comparison here. Ant seems to have far more data about users so that its credit score is crazy accurate.

Hard to believe this story. Feel like the police department is not working as expected.

Thanks for the explanation: now it seems to me the comparison is not perfect but still meaningful.

> The tools used for benchmarking were BenchmarkTools.jl for Julia, microbenchmark for R, and timeit for Python.

Is this a meaningful comparison? These benchmarks repeatedly call the function (with same input file) many times and then take the average of the time spent. However, in real world, we only read a specific file ONCE -- if we call the csv reader several times, it is almost always for different files.

In Python "a" or "b" returns "a".

If you read the dict union operator "|" as something similar to "or", (like me), you are in for a surprise:

>>> x = {"key1": "value1 from x", "key2": "value2 from x"}

>>> y = {"key2": "value2 from y", "key3": "value3 from y"}

>>> x | y

{'key1': 'value1 from x', 'key2': 'value2 from y', 'key3': 'value3 from y'}

What I expect is

>>> x | y

{"key1": "value1 from x", "key2": "value2 from x", "key3": "value3 from y"}

MSFT stock rise about 5.62% (or $100b), likely due to the news to take over TikTok in US. If this is a fair trade, one would not expect MSFT stock to increase.

There is limited eye time. TikTok is on the path to winner-takes-all in the space it is currently in. I would value it at $300B given its huge potential that none of its competitors or copiers can have. $50B is low imo -- likely caused by the political pressure.

same here.

One of my friends has type I diabetes, in more than one times, she needed urgent visit/call to her doctors just to get the medicine she has being taken for years.

Finally we come to common sense.

I like Unity and did not have good impressions of GNOME.

Any place to download the data used for the regression model?

"For each 1 percent of all residential units in a neighborhood listed on Airbnb, rents in that neighborhood went up 1.58 percent, Stringer said."

Correlation does not imply causation[1]. Looks to me this is one of the studies that I would label as "shallow" or "problematic".

https://en.wikipedia.org/wiki/Correlation_does_not_imply_cau...

Is it functional as of today? The perception directory is almost empty (only skeleton).

For all the cases I worked on, R is not slow at all. At home, I wrote a R package to do deep learning (include most common layers, but only conv layer was implemented in C++) from scratch, and I also wrote one in python (using numpy + numba). My R version is same fast as python on the MNIST dataset (without conv layers, hence all code is in plain R or python).

Does anyone know when Julia 1.0 will be out?

Good point. On the other hand, the confidence interval I provided is correct as long as the original survey and the repeat are sampled in the same way (regardless whether it is un-biased sampling or not)

1000 is enough for this case.

Depends on how "accurate" you want to measure a ratio. Here the ratio is the percentage of Americans that support net neutrality.

In this case, sample size is 1000, estimated ratio is 76%, the 95% confidence interval is 76% +/- 2.65%, which means if you repeat this survey again, you have a 95% chance that new estimated ratio is within 76% +/- 2.65%.

Edit: 99% confidence interval is 76% +/- 3.48%

per wikipedia

Money laundering is the process of transforming the profits of crime and corruption into ostensibly 'legitimate' assets.