HN user
If the confidence interval width is 2 * 14.0%, how are you detecting a statistically significant difference between 58% and 50%?
The 95% CIs on both timeseries pretty much always cover the baseline number, which is not consistent with the result being statistically significant.
The process where resources accrue to those with more resources is called the Matthew Effect. It explains, amongst other things, why the degree distribution of social networks follows a power law.
There's a nice experimental test of this where showing the number of previous downloads a song has makes it more likely to be downloaded (but not to the extent that it entirely overrides the quality of the song. <https://www.princeton.edu/~mjs3/salganik_dodds_watts06_full....>
Examples include converting boxplots into violins or vice versa, turning a line plot into a heatmap, plotting a density estimate instead of a histogram, performing a computation on ranked data values instead of raw data values, and so on.
Most of this is not about Python, it’s about matplotlib. If you want the admittedly very thoughtful design of ggplot in Python, use plotnine
I would consider the R code to be slightly easier to read (notice how many quotes and brackets the Python code needs)
This isn’t about Python, it’s about the tidyverse. The reason you can use this simpler syntax in R is because it’s non-standard-evaluation allows packages to extend the syntax in a way Python does not expose: http://adv-r.had.co.nz/Computing-on-the-language.html
This specific analysis isn’t p-hacking because although they conduct multiple tests, they report all of them rather than just the statistically significant ones.
They should however account for multiple testing. The Bonferroni correction (which is conservative) would set the alpha level to 0.05/5=0.01, for which the 1 day after result is still (just) statistically significant.
Not to say there couldn’t be other problems.
Thread is called “fils” in French, meaning “son”, considering channels are parents of threads in a sense.
Could be, but “fil” also literally means “thread” in the sewing sense.
You need assumptions exactly for the things which cannot be tested against reality. Otherwise why assume rather than measure?
all of
designed a dozen sites I never published I was never on time I couldn't get up for class on time
And depending on why
don't drive I feel like an alien, and most everyone drives me insane. most of them I cut off without a word, and those that reach out I resent
also sounds like how someone with ADHD could describe themselves, and the other issues could be downstream from that
Can buy-and-hold investors use margin?
This study can't see past its own midwit view that there is an objective “detailed, literal” reading that necessarily produces the same interpretation of the text that the authors have.
The students in the study are responding in a rational way to the way HS English is taught: the pretense is that you're deriving meaning/themes/symbolism from the text, but these interpretations are often totally made-up[^0] to the extent that authors can't answer the standardized tests about their own work[^1]. The real task is then to flatter the teacher/professor/test-setter's preconceptions about the work — and if the goal is to guess some external source's perspective, why shouldn't that external source be SparkNotes?
This ambivalent literalism is evident in the paper itself: - one student is criticized for "imagin[ing] dinosaurs lumbering around London", because the authors think this language is obviously "figurative". But it's totally plausible that Dickens was a notch more literal than only describing the mud as prehistoric! In the mid-1850s the first descriptions and statues of dinosaurs were being produced, there was a common theory that prehistoric lizards were as developed as present mammals, so maybe he's referring to (or making fun of) that idea? - the authors criticize readers for relying on SparkNotes instead of looking up individual words in the dictionary. But "Chancery" has ~8 definitions, only one of which is about a court and "advocate" has ~4. Is it more competent to guess which of those 32 combinations is correct, or to look up the meaning of the whole passage instead? There's whole texts dedicated to explaining other texts, especially old ones — does pulling from those make you a bad reader? - they say that a student only locates the fog vaguely rather than seeing that "it moves throughout the shipyards". That's not in the text though: the fog is only described as moving laterally in two of the locations, and never between different parts of the yard. Maybe the fog is instead being generated in each ship and by each person, as is the confusion in the High Court of Chancery? (More pedantically still: are all these boats just being built? If not, wouldn't they be at docks or wharfs rather than shipyards?)
I think the underlying implicit belief is that there is always one correct interpretation of the text, at one exactly correct level of literalness, derivable from only the text itself. But by the points students are in college they will have been continuously rebuffed for attempting literal interpretations that don't produce the required result, and unsurprisingly they end up unsure which parts of understanding are mechanical and which are imaginative.
[^0] https://www.theparisreview.org/blog/2011/12/05/document-the-... [^1] https://www.huffpost.com/entry/standardized-tests-are-so-bad...
I think this (jokingly) assumes that PhDs become academics, which is not always true. Also, PhDs are generally free and if you leave early you get a Master's anyway. Terminal Master's you have to pay for.
Let's say that The Rebel Times has a headline "Member of the Imperial Senate on a diplomatic mission boarded and arrested without cause" while the Empire Daily reports that "Leia Organa, part of the Rebel Alliance and a traitor, taken into custody". Following your process, the "what" is just that Leia was arrested.
Then, the Rebel Times says "Moisture farmer with magic powers joins fight against Empire", but the Empire Daily has "Moisture farmer joins fight against Empire". the common whats are just that a moisture farmer joined the Rebel Alliance, which is true, but much less consequential than if he had magic powers.
Later, the Rebel Times says "Secret Empire super-weapon destroyed at the Battle of Yavin", and the Empire Daily publishes... nothing because they don't want to admit defeat. There's no common information between these stories (because there is no second story), so looking for common whats would conclude that nothing happened.
If the process of analysing the news accounted for the fact that the different outlets are interested in presenting different whats, it could conclude that the fact that the Empire Daily published nothing about the third story doesn't mean that it didn't happen. In the second case, if it could account for the Empire wanting to suppress information about the Force, the conclusion would be that Luke joining the Alliance is somewhat more of a big deal than otherwise. Even in the first case, it might realise that the fact that the two sources don't agree about Leia doesn't mean that one side isn't right.
I have no doubt the same difficulties would apply to - say - planting and harvesting crops on a schedule
Not necessarily. Before the Industrial Revolution artisans were paid by the piece and peasants harvested between the crop being ripe and it being damaged by the weather. <https://dhayton.haverford.edu/wp-content/uploads/2021/10/Tho...>
Both of those have immediate rather than delayed (artificial?) consequences, which might have been easier to handle.
When factories tried paying peasants by the hour, they would work as much as they felt like (e.g. not showing up at all on "Saint" Monday) until they were taught time-discipline.
In case anyone else is a fan of these unsual old displays, you might enjoy Posy's passionate videos which are a unique combination of curiosity, macro/slowmo shots and humor. Especially relevant is this one which is about what seems to be a very similar (but different sized) red/green/blue Casio CSF-7950 LCD: https://www.youtube.com/watch?v=jLew3Dd3IBA
Also as of more recently, https://en.wikipedia.org/wiki/Claudia_Sheinbaum
I think this is too strong. If art has no inherent meaning, you can't say that it's not making a political statement, only that whether it's making a political statment or not is undefined.
Otherwise, saying that something is not making a political statement is actually a very strong political statement in itself (e.g. things which are "not political" can't be negotiated or changed).
Too easy:
Van Gogh painted it in an asylum (social safety net) run by Franciscans (repeal of anti-Church laws during the Third Republic) because he had self-admitted after cutting off his own ear (treatment of physical violence as a medical problem) because he owed money to Gaugin (private property). It includes an imaginary house and was painted from memory because he wasn't allowed to paint in his room (medical treatment in a total institution), but he was allowed to use a spare room in the half-full asylum which was normally for the wealthy (obvious). That said, he considered it a painting from nature, rather than an "abstraction", a form which was preferred by Gaugin in order to indicate harmony between man and nature (a fin de siècle concern which also directly lead to the rejections of liberal democracy and bourgeois society in the early 20th Century). On his own account, the subject matter of stars connotes a spiritual hope — and you can make what you will of the more specific religous and astronomical interpratations that came later.
In a narrow sense, there's no political art except electoral propaganda, in a broad sense anything made by humans is political.
So I don't find the game of definitions is not particularly interesting. It's not even that interesting whether the political content of art is intentional or self-concious on the part of the author or whether it's imposed by others. The process which I think is worth thinking about is the very second-order process we're engaging in now: who wants to make a claim about the politics (or lack thereof) of art, and why are they interested in doing that?
the US educational system is generally so poor that it's not generating enough people with the necessary technical capabilities to meet employer demand
The US educational system might be generating enough people with those skills, but a very large proportion of them are foreigners who would need visas to work (40-50% of STEM postgrads: https://crsreports.congress.gov/product/pdf/IF/IF11347)
"in the grim dark future there is only war"
Unalienable IDs might be the default of small-scale face-to-face societies, but not ones which operate at scale. Not even "Bob" can serve as a fixed identifier without intentional state action — (sur)names were very often intentionally assigned to make administration (e.g. taxation) easier:
Campaigns to assign permanent patronyms have typically taken place, as one might expect, in the context of a state’s exertions to put its fiscal system on a sounder and more lucrative footing. Fearing, with good reason, that an effort to enumerate and register them could be a prelude to some new tax burden or conscription, local officials and the population at large often resisted such campaigns.
https://theanarchistlibrary.org/library/james-c-scott-seeing...
The underscore _ can be used as a throwaway variable to discard unwanted values:
So can any other variable, using underscore is just a convention to make it obvious that you're not planning to re-use it (it doesn't get GCed more aggressively or anything).
Similarly, private methods being prefixed with an underscore is also just a convention, you can access them from anywhere.
However, double underscores are used for magic attributes and name mangling for class attributes, which are interpreted differently! (See: https://stackoverflow.com/a/1301369)
It looks from Figure 3 that they fit a sinusoidal curve to the data, and then fit an exponential to the residuals of *that* fit. That exponential shows a sharp increase from hour -2, but the curve is not a good fit to the rest of the data, and it almost has to show an increase somewhere because of the functional form.
I would have assumed the way to evaluate this method would be to back-test it: figure out when in the past this would have predicted that there would be an earthquake, and measure the accuracy/precision/etc. of that prediction.
It often surprises me how often disciplines don't have take a predictive approach by convention: if you make predictions at least it's possible to be *wrong*.
In August 2022, I started 17aE2, non feminizing estradiol (Rx), based upon male longevity evidence. My dosing protocol is 4mg wk, transdermal. Since we started, there’s been no change to serum estrogen or testosterone and no known side effects.
Source: https://blueprint.bryanjohnson.co/#diet-and-supplements
It doesn't count `failure` — just the number of rows. But neither does the pandas version: `pd_df.groupby(['date'])['failure'].count()` and `pd_df.groupby(['date']).count()` are the same except the former returns a single `pd.Series` with the count and the latter produces a `pd.DataFrame` where each column has the same count (not super useful).
e.g.
> iris.groupby('species').count()
sepal_length sepal_width petal_length petal_width
species
setosa 50 50 50 50
versicolor 50 50 50 50
virginica 50 50 50 50
> iris.groupby('species')['sepal_length'].count()
species
setosa 50
versicolor 50
virginica 50The ergonomics of grouping and aggregation in R are really much better because libraries can make of its non-standard evaluation[^0] (which in other cases also makes the language a nightmare to deal with).
Compare:
pd_df.groupby(['date'])['failure'].count() # pandas
pl_df.groupby(pl.col('date')).agg(pl.count('failure')) # polars
dt[, .N, date] # R data.table
The reality of generating diagrams from code is that although you’ll generate diagrams that are a 1:1 accurate representation of your infrastructure and source code, they likely won’t be very useful for your team
See also the Borges short story On Exactitude in Science:
...In that Empire, the Art of Cartography attained such Perfection that the map of a single Province occupied the entirety of a City, and the map of the Empire, the entirety of a Province. In time, those Unconscionable Maps no longer satisfied, and the Cartographers Guilds struck a Map of the Empire whose size was that of the Empire, and which coincided point for point with it. The following Generations, who were not so fond of the Study of Cartography as their Forebears had been, saw that that vast Map was Useless, and not without some Pitilessness was it, that they delivered it up to the Inclemencies of Sun and Winters. In the Deserts of the West, still today, there are Tattered Ruins of that Map, inhabited by Animals and Beggars; in all the Land there is no other Relic of the Disciplines of Geography. [^0]
More problematically, it's generally the case that more than one abstraction is possible and the available evidence can't on its own decide which is better.[^1]
(This title should probably be hyphenated as in the original source: "Model-Code Gap" reads as in "gap between model and code" as opposed to "the gap which can be found in code which could serve as a model".)
[^0]: https://kwarc.info/teaching/TDM/Borges.pdf [^1]: https://plato.stanford.edu/entries/scientific-underdetermina...
In the U.K., owners of televisions receiving broadcast signals (and now VoD) are required to pay for an annual license. The enforcement of this is mostly through sending letters, but there are also TV Detector Vans [0] which work in a similar way to TEMPEST — or at least, that's allegedly how they work because there's some debate about whether they really exist or if they're just a PR tactic.
It is crucial for data scientists to carefully consider the quality and representativeness of the data they are working with
To bring this fully to the domain of epistemology, you could go further: why are certain markers of quality or measures of representativeness valid? If we develop a measure and it gives us a particular answer, can we tell whether the data or the measure is at fault.[1]
[statistical significance] refers to the likelihood that a result or relationship observed in a sample is not simply due to chance, but rather reflects a genuine trend or pattern in the population.
This is in general not correct — statistical significance is about the probability of an observation assuming that the null hypothesis is true.[1] There is a narrow context in which you could interpret a p-value as being about the probability that there is a true effect: if you're doing Bayesian inference with a flat prior, but that is itself a very strong claim to make.
[0] https://plato.stanford.edu/entries/measurement-science/#TheL... [1]: https://en.wikipedia.org/wiki/Misuse_of_p-values#Clarificati...
The first large failed project was a request for us to attribute all forecasts to fundamental business drivers, and automatically explain the reason for all forecast misses.
This is described as an "Insane Business Request", but it seems like a useful business question to be able to answer, although maybe scientifically difficult.
Most scientific problems I’ve worked on could be solved by a correct representation of an empirical distribution in a histogram. The next largest group needed a linear regression. The final group needed a statistical model from the first chapter of a PhD textbook.
A mentor once told me that the best way to get started in a new modelling field is to read all the cutting edge papers, and then actually implement whatever baseline model they're claiming to improve upon.
This is like Cartiesian scepticism [0] justified by attempting to reject the fixed points of the prior in Bayes' rule [1]. The implication is that your belief in something should always be an epsilon away from 100% (and presumably 0%).
[0] https://plato.stanford.edu/entries/skepticism/#ArguForCartSk... [1]: https://www.austintripp.ca/blog/2018/08/21/bayes-no-certain-...