boomboompsh
boomboompsh
data and stuff I guess
8 posts
I have made literally two posts both about data. I guess that's what this blog is.
Don't wanna be here? Send us removal request.
boomboompsh · 6 months ago
Text
Foone is overstating the theoretical possibility of danger. Noone has been able to reliably detect *any* x-rays from CRTs. From international labor organization guidance (they call them VDUs for "visual display units"):
Numerous radiation measurements both in field and laboratory conditions have been conducted worldwide in attempts to detect ionizing (X-ray) radiation emissions from VDUs. Basically, these attempts have failed, in that no detectable emissions beyond the "natural" or "instrumental" backgrounds could be detected (Cox, 1984; Moss et al., 1977; Weiss and Petersen, 1979; Phillips, 1981; Terrana et al., 1982; Wolbarsht et al., 1980; Paulsson et al., 1984; Murray et al., 1981; Health and Welfare Canada, 1983; Pomroy and Noel, 1984; Joyner et al., 1984; Bureau of Radiological Health, 1981). Emissions of X-rays from VDUs are so weak that they cannot penetrate the front glass screen and so cannot be detected against the normally encountered background levels of ionizing radiation.
https://www.ilo.org/publications/visual-display-units-radiation-protection-guidance pages 10-11
It goes on to state that some of these measurements showed no detectable levels even after breaking parts of the TV to try to increase the x-ray output. A few TVs from before 1978 produced measurable levels and were pulled from the market or never made it to market and xrays haven't been detected since. There is no danger from the ionizing radiation.
Hey are those bulky old tvs dangerous to be around or is that just something my mom told me so I wouldn't sit too close to the screen
4K notes · View notes
boomboompsh · 9 months ago
Text
had to read through the whole article and see the title again before I realized that "stardom" refers to the state of being a star and not an obscure kind of dominatrix.
stardom dreams, stalking devices and the secret conglomerate selling both
over the last half a year, @rhinozzryan and i have worked on an investigation into Tracki, a "world leader in GPS tracking", and ExploreTalent, one of the biggest talent listing services in the world. what the hell do those two have in common?
Tumblr media
(feature art by @catmask)
7K notes · View notes
boomboompsh · 2 years ago
Text
Tumblr media
5 posts!
Yuuyyyuu u.
-my pocket
0 notes
boomboompsh · 2 years ago
Text
According to this study:
the conclusion of which is the title, age of pubertal onset does not affect final height, at least in the normal range of ages, noone in the study had puberty onset past age 14. Individuals with delayed pubertal onset may be shorter on average, though the effect is not observed in all studies in this review:
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8579478/
I have been unable to find any studies on height in individuals with mais, though I did find a study that found the heights of pais and cais individuals to be inbetween typical male and female heights.
https://pubmed.ncbi.nlm.nih.gov/35629193/
What's funny is I am that type of intersex where you have a delayed puberty and less masculine traits and I'm still tall as hell
448 notes · View notes
boomboompsh · 2 years ago
Text
Tumblr media Tumblr media Tumblr media Tumblr media Tumblr media
423K notes · View notes
boomboompsh · 2 years ago
Text
Tumblr media
5K notes · View notes
boomboompsh · 2 years ago
Text
I did a similar study to sprinklecipher and acheived some different results. sprinklecipher limited the number of results by filtering recent videos. since recent videos necessarily will be biased towards lower views, the incidence of one-view videos was severely undercounted.
Method (short version):
I searched youtube videos using random strings of alphanumeric characters, giving me hits on videos that have those strings in their video ids, which are assigned randomly,so there shouldn't be any relevant bias once videos that hit on different parts of the metadata were filtered out.
Results:
TL;DR: zero is slightly more common ~52-59% (95% CI) compared to one view.
Here are the counts from the bottom end of the histogram of youtube views I generated. 2 was a surprisingly strong contender. the data was generated from a total sample size of 6142 videos
Tumblr media
From the differences between this data and sprinklecipher's, it appears the "watching the video to make sure it uploaded correctly" is not responsible for most 1-view videos. It would appear that in the sample sprinklecipher took, people had just trusted youtube to work, and only later would those videos ever receive a view. I might make a followup post where I go into detail about the code I wrote to take this data and potentially other interesting things about this mostly unbiased sample of youtube videos I have taken. Dealing with the youtube data api was kinda frustrating and I want to vent about that.
Here is a link to the spreadsheet where I did all the data processing:
(I made this post late at night and did not realize file.io deletes uploads after they are downloaded once. This link will last thirty days.)
There are nine video id rows that are blank. Those are from a few videos I manually recorded the view counts of my first day. I did bias the results slightly by taking a new sample every day and only publishing once my results became significant. I am not worrying about this too much for a fun problem, but it is an important bias to worry about in other applications.
@sprinklecipher here is my data, I didn't want to publish when I initially responded to your blog post because I had an n of 9. literally the only useful conclusion I could reach back then was that the ratio wasn't 197:1.
854 notes · View notes
boomboompsh · 2 years ago
Text
I have been looking for data on this as well and trying to avoid bias as much as possible. I did this by searching (using the youtube api) for strings of random alphanumeric characters and checking the view counts of the results manually (ran out of quota). This finds videos that have those strings in their video ids, which should be unbiased as video ids seem random. I've been getting roughly equal numbers of 1 and zero view videos and the probability of that happening if there were 197 times as many zero-view videos as one-view as it seems from your data is <0.00000001 according to a binomial test calculator I found. It seems like youtube must be promoting zero-view videos in search results.
854 notes · View notes