HN user
This blog post went up on Saturday, September 12. It says:
> MUCH cleaner air will push in by 4 PM Sunday over the coastal zone and will just reach Seattle late in the afternoon
> By 1 AM Monday, air quality will be hugely better in western Washington
As of now (Tuesday) there is no sign of clearing. Here in Seattle we are still in the "very unhealthy" range for AQI.
From that CityObservatory article:
Zumper has also made a name for itself through its “National Rent Reports”
—more or less monthly press releases that claim to track median rental prices around the country.
These reports have received copious media coverage, from the Bay Area to Seattle to Nashville to Chicago to Boston to LA to Miami to Denver, and so on.Note that Zumper is not a reliable source for median rent data. CityObservatory wrote an article about their data problems a few years ago [1]. Zumper's data is, of course, based on apartments that are for rent and doesn't include currently occupied units. That alone skews high, especially when there is a lot of higher-rent new construction hitting the market. Also, it looks like Zumper's data skews towards higher-end neighborhoods.
For a broader look at the rental market, including occupied units and rent-controlled units, you could just consult ACS data. That says that median rent for all occupied 1-bedrooms in San Fransisco was $1912 in 2017 [2].
[1] http://cityobservatory.org/journalists-should-be-wary-of-med...
[2] https://censusreporter.org/data/table/?table=B25031&geo_ids=...
I've heard good things about oshpark[1] for hobbyists, but I personally have never used them.
Its striking how much less bunching there is on the two routes (P/G) that operate in exclusive right-of-way.
The CO2 is also an input; it originally comes from the environment, so this is net-zero emissions in the same way that biofuels are. From the article:
Although the bus emits CO2, Team Fast argues that the original CO2 used to create the hydrozine is
taken from existing sources, such as air or exhaust fumes, so that no additional CO2 is produced
- it's a closed carbon cycle in the jargon.Jed Kolko, former chief economist at Trulia, thinks stories like this are either exaggerated or wrong. His basic claim is that urban revival is limited to childless professionals in their peak earning years. See [1], or any of his posts at [2] for the data and analysis.
[1] http://jedkolko.com/2016/03/25/neighborhood-data-show-that-u...
Well, supposing this is correct...Congratulations to Anthony and the rest of the Kaggle team! Those guys do a great job. Hopefully they get rewarded for it.
This listing seems to be missing https://upverter.com/.
I use both R and python quite a bit. I prefer python as a programming language. Here's my take on 'Why learn R?': (1) R/ggplot is hands-down better for plotting than anything in python. I also think that R is better for EDA generally. (2) Many smart, knowledgable people use R and publish their code. To learn from it, you need to know enough R to read and modify it. (3) R has better package support than python in several common data analysis domains. For example, in forecasting and in graph analysis, the best R packages available are much better than the best python packages.
>“It’s a paradox,” said Valentin Bote, head of research in Spain at Randstad, a recruitment agency. “The unemployment rate is too high. Yet we’re seeing some tension in the labor market because unemployed people don’t have the skills employers demand.
There's no real paradox there. Employment of young people, and therefore normal career progression for that cohort, essentially shut down for 6-8 years. Now the pipeline is a little empty.
A lot of people are going to dump on this billion dollar valuation, but we don't have the terms. Presumably there is liquidation preference. According to Crunchbase [1], the total capital raised is about $160 million after this money, so the investors don't need this company to be worth $1 billion to come out fine. Sam Altman made the point a few days ago that a lot of these late-stage private financings are kind of debt-like [2, about paragraph 8-9 and footnote #2], with the result that the valuations aren't really meaningful.
I wouldn't be so down on vocational training for software...It is a field that has an unusually high requirement for ongoing learning. And when you have to learn some new tech, usually you have to do it now, and in-place, not next September, in the nearest college town. Given the high ongoing learning requirements that software has, I think there is a pretty clear need for training that's delivered where you are and when you need it, and traditional schools won't be able to do that.
>I'm sorry, continuing resolutions are now our ideal?
I did not say that continuing resolutions are an ideal. My point was that although budgets are complex, only a relative small part of them changes from one cycle to the next. People don't have to digest the whole budget, because they already know what was in it before. They only have to understand the changes.
The ACA wasn't negotiated in secret. Budgets are normally continuing resolutions; they only negotiate the delta from previous years, which is usually small. And again, budget negotiations don't involve years of secret negotiations.
How are people who are starting from zero on this supposed to understand it in 60 days? This disclosure only a little more than a transparency fig leaf.
The point about the late-stage investments being not-really-equity is a great point. So my question is: why do financial journalists just about always miss it when they write about tech unicorns?
In the post, PG states that First Round's study is evidence of gender bias in VC financing. But footnote [2] is important: Uber was excluded as an outlier. Now...excluding Uber is reasonable (it is sort of an outlier), but so is not excluding it (it was a company that First Round invested in). When the conclusion from a data analysis depends on which way you go on something like this - which of two reasonable alternatives you pick - then the results are fragile and they don't really support either conclusion very well.
I've played around with this some, and the recognition isn't perfect, but I'm very impressed with how well it can pick up a new pattern from even just one example.
It seems like they should really call that top-line, $10 bn, number 'bookings' and the second, $2 bn, number 'revenue'.
Fair enough... but the issue here is about findings of an effect versus no effect. It's about statistical significance. In order for single-comparison p-values and the like to be valid, there has to be a (one) single comparison. There is a way to do 'any-of-k' testing, but the required effect sizes get larger.
It seems that the main issue here is that with pre-registration, study authors have to pick a single measure of primary benefit at the outset, whereas before, they might have made that choice after getting results back. The original study is at PLoSONe, and it is not a difficult read[1]. From that source:
>Prior to 2000, investigators had a greater opportunity to measure a range of variables and to select the most successful outcomes when reporting their results... Among the 25 preregistered trials published in 2000 or later, 12 reported significant, positive effects for cardiovascular-related variables other than the primary outcome.
That is, in most cases, there are large effects for some outcome, and if they get to choose the primary outcome after looking at some results, they could have been cherry-picking the outcome variables.
[1] http://journals.plos.org/plosone/article?id=10.1371/journal....
Is this dataset available anywhere? I see the code on github, but I don't see the data anywhere.
The blog post seems to be getting modified at this moment. When I first saw it, it didn't have anything about the misclassifications, but that has been added now.
I've only glanced at her code, but it looks like[1] the predictions are from held-out data.
EDIT: All of the data was used in forming the PCA basis, but that isn't (necessarily) an error, depending on the use-case. And the logistic regression model was evaluated on held-out data.
[1]https://github.com/graceavery/Eigenstyle/blob/master/visuals...
Scroll down to where it says:
It even works for dresses that were not in the training set:
Nonlinear methods can result in a smaller (lower-dimensional) representation, but they are non-linear so they are harder to use and usually require more data. On the other hand, PCA is easy and having 50 or more principal components is often not a problem, unless you are doing visualization. With the additional representational capacity from just keeping extra dimensions you can still get good reconstructions.
> Was there image processing to distill the dress into major components, hand tagging of variables (like short/long, color, etc), or just raw pixel data processed?
Its pretty clear that this is PCA on raw pixels of the images, so no, there is no hand-tagging or anything else. The eigendresses are just the eigenvectors from PCA projected back into the original, image-scale, basis.
It does survive, in a certain sense, in scientific programming and data science. Both iPython notebooks and Rmarkdown are a sort of literate programming, although with the emphasis on the text more than the code. In that setting, the executable artifact is not really more important than the explanation of why the code does what it does, so the extra overhead is justifiable.
Rmarkdown example: http://kbroman.org/knitr_knutshell/pages/Rmarkdown.html
iPython notebook example: http://nbviewer.ipython.org/github/empet/Math/blob/master/Do...
Ok...That's pretty cool. Thx :)