Mind the supplemental data... itâs a mess
seen from TĂŒrkiye
seen from China
seen from United States
seen from United States

seen from United States

seen from Israel

seen from United States
seen from Saudi Arabia
seen from United States
seen from China
seen from United Kingdom
seen from China
seen from Guatemala
seen from Chile

seen from Australia
seen from China
seen from United States

seen from Malaysia

seen from United States

seen from United States
Mind the supplemental data... itâs a mess

Anya is live and ready to show you everything. Watch her strip, dance, and perform exclusive shows just for you. Interact in real-time and make your fantasies come true.
Free to watch âą No registration required âą HD streaming
So You Thought Hong Kongâs Pollution Comes From the Mainland? â Supplemental Data
This article includes supplemental data for the main article âSo You Thought Hong Kongâs Pollution Comes From the Mainland?â
This supplemental data will show how high and low wind speeds affect the PM2.5 levels and hence Hong Kongâs pollution. Ultimately, this supplemental article draws the same conclusion as the main article.
 Does Wind Speed Affect PM2.5?
What if particulate matter can only be transported at high speeds? In order to test if wind speed affects PM2.5, I repeated the analysis for PM2.5 data at wind speeds below average and above average.
The data shows that lower wind speeds in Hong Kong almost always correlated with lower PM2.5 levels, regardless of the wind direction.
However, this also shows that air pollution was significant even without strong winds blowing PM2.5 from Mainland China. This means that there are PM2.5 sources within Hong Kong as well.
When wind was blowing at higher speeds from the direction of Mainland China, the PM2.5 level was about 20% higher than average.
Get the latest clean air tips!
Get updates on masks, air purifiers and air quality delivered straight to your inbox.
This Data Explains Why You Should Never Use Your Purifierâs Auto Mode - Supplemental Data
Data and Test Details
 Test Methods: These tests were identical to my previously published tests, details of which can befound here, here, and here.
Original Data: The original data for all 13 tests is available as an Excel file here
 Average Outdoor PM2.5 During Tests
Itâs important to note the average outdoor PM2.5 during tests because purifiers achieve smaller percentage reductions when outdoor air is really bad. According to the US Embassy, outdoor PM2.5 was 128.2 micrograms during the Xiaomi tests and 108.4 during the Philips tests. The tests were done in the winter, so these averages about normal for Beijing.
 Theoretically, if the auto mode is working as itâs supposed to, outdoor air pollution shouldnât matter. If air is bad, the machine should just spend more time in high mode.
 Fluctuations in Outdoor Pollution
If outdoor pollution gets a lot worse during the test, that could make it seem like the purifier isnât working well (although again, if auto mode is working like it should, that shouldnât matter, at least pollution exceeds the ability of the high setting). But to be conservative, I re-analyzed the data with fluctuations in mind.
 During the Philips tests, outdoor air was remarkably stable. The only large fluctuations worked in the Philipsâs favor (falling during the test). During those two days, the Philips averaged 50.9% reductions in 0.5 micron particles, so no big difference.
 During the Xiaomi tests, outdoor air varied a lot. Outdoor PM2.5 fell on some days and rose on some days. The fluctuations didnât seem to matter much. For example, on the most extreme day, outdoor air rocketed from 178 micrograms to 467, yet the percentage reduction was right near average (65% versus 61%).
 In sum, with the auto mode tests, I think outdoor fluctuations are less important than when the purifier is on a constant setting. Instead, a lot rides on whether the auto mode has decided to turn on at the moment the Dylos takes a reading. That introduces randomness into the results, although with enough tests, the randomness should smooth out.
 Converting Dylos 0.5 Microns to PM2.5: As in previous tests, I converted the Dylos 0.5 micron counts to PM2.5 ug/m3 by dividing them by 100. In prior tests, this semi-official formula from Dylos worked pretty well.
 Can you really not keep the Xiaomi 2 on high setting?
Apart from the tests we have done, we went to Xiaomi directly just to confirm on this. Hereâs what they say:
 Translation:
Smart Air: Is it not possible to keep the purifier on Favourite mode?
Xiaomi: No, itâs not possible
Smart Air: I heard a friend say you can use the Xiaomi App to keep the purifier running on Favourite mode.
Xiaomi: Itâs not possible to keep the purifier constantly running on favourite mode
Smart Air: OK, thank you.
Xiaomi Auto Mode Leaves Air Unsafe for 86% of Hours - Supplemental Data
This article includes supplemental data for the main article âXiaomi Mi 2 Auto Mode Leaves Air Unsafe for 86% of hours | Reviewâ
Data and Test Details
Settings
The app asks what size the room is. How Xiaomi uses that data is a little opaque, but it would be logical that settings for larger rooms will run the purifier harder, so we used the highest allowable setting, 34-37m2.
It should be noted that the Xiaomi also has a turbo mode, which Iâve heard people describe as âsounding like a jet taking off.â This really isnât meant to be used while people are in the room. Instead, itâs meant to be used temporarily before returning home.
Tests in a Different Room
Weâve used this same bedroom in Chaoyangmen, Beijing for many of our testsâof the Original DIY, Cannon, Philips, Blue Air, and IQ Air. So I doubt thereâs something weird about this room that somehow hurts the Xiaomi but not other machines.
But you never know! So to be sure, we conducted an overnight test in the Smart Air office near Sanlitun. The results are shown in the noise test above. They show the same pattern as the other room tests. Thus, the results donât seem to be something weird happening with that particular test room.
Tests with a Different Particle Counter
Weâve used the Dylos Pro for most of our tests, so this is constant across tests. But particle counters can break or lose accuracy over time. Thus, we carried out the noise/particle test above with a different Dylos Pro. The fact that the pattern of the results is the same suggests the results are not because of any problems with the particle counter. (That conclusion is also hinted at by the fact that the Xiaomi performed well for the first three hours while on high.)
How Bad Was Outdoor Air?
Itâs important to analyze how bad outdoor air was during the tests because my analyses have shown itâs harder to achieve a high percentage reduction when outdoor air pollution is bad. (Or put another way, itâs easier to achieve a high percentage reduction on relatively clean days.) For example, here is the relationship between effectiveness and outdoor air pollution for the Blue Air 203:
However, theoretically, the Xiaomi auto mode shouldnât be affected by outdoor air pollution. If it can accurately detect air pollution levels and turn on the fan in response, the results shouldnât be affected by outdoor air pollution levels (until we get to levels that are too high for even constant high mode to clean). But to be conservative, I analyzed hourly outdoor air pollution data from the US Embassy, about 7 kilometers from Annaâs home.
During the auto mode tests, outdoor air averaged 128 micrograms. During the high-mode tests, outdoor air averaged 246 micrograms. Both of these are higher than the average in Beijing (90-100 micrograms). So itâs worth seeing how the Xiaomi did on days with lower pollution.
I analyzed the four days with lowest outdoor air pollution (average 89 micrograms; 11/29, 12/2, 12/5, 12/7). On these days, the Xiaomi averaged a 56% reduction in 0.5 micron particles and 87% in 2.5 micron particles. Thus, the poor results in the main tests do not seem to be because of outdoor air pollution levels.
Outdoor Air Fluctuations
Besides the baseline level of outdoor air pollution, itâs also important in these real-world tests to look for large fluctuations in outdoor air pollution. If outdoor air gets a lot worse during the test, it can look like the purifier is not cleaning the air very well. And on the flip side, if outdoor air pollution goes down a lot during the test, it can look like the purifier did a great job.
First, I analyzed the data after removing any test day where outdoor air fluctuated more than 100 micrograms from beginning to end (4 days total). That left 8 test days. The result was very similar: on these stable days, the Xiaomi average a 67% reduction in 0.5 micron particles and 86% in 2.5 micron particles.
Next, I analyzed all days where outdoor PM2.5 micrograms changed no more than 50% from baseline at any point during the test. This âat any pointâ criterion is more stringent, and it left three test days (12/11, 12/12, 12/18). Again, the results were nearly identical: a 69% reduction in 0.5 micron particles and 83% in 2.5 microns. In sum, the poor results did not seem to be caused by fluctuations in outdoor air pollution.
% Hours of Unsafe Air
To calculate the percentage of unsafe air for the Blue Air 203/270E (3,600 RMB), and Philips AC4072 (3,000 RMB), I used the data in my previously published tests. For the Cannon, I used three series of tests: my original tests, tests where tested whether adding a pre-filter affects performance (below), and tests comparing performance after adding a carbon layer.
That gives a lot of data! Hereâs exactly how many hours of data I had:
Purifier
# Hours of Unsafe Air (>25ug)
Average PM 2.5 During Tests (ug/m3)
Phillips
45 78
Blue Air
41
113
Xiaomi
93
209
Cannon
98
111
IQ Air 80
146
% Unsafe Hours: How bad was outdoor air?
Itâs important to compare just how bad outdoor air was during the tests. If one machine has lots of hours where air was just above 25 micrograms, it would be a lot easier for the machine to clean the air. So, I calculated the average outdoor micrograms for all of these unsafe hours. For reference, Beijingâs air has averaged about 90-100 micrograms for the last 7 years according to the US Embassy.
In the chart above, I analyze what the average outdoor PM2.5 concentration was for the different tests. The Xiaomi really stands out, and that could unfairly lower its results. Thus, I re-analyzed data only looking at days with lower outdoor air pollution, with an average of 138 microgramsâlower than the IQ Air. The result was similar to the original analysis: 83% of hours were unsafe.
To be even more conservative, I analyzed the two days with the lowest outdoor concentration, averaging 108 micrograms. That is lower than all of the other machines except the Philips. On these days, 68% of hours were unsafe. Thus, the Xiaomi was leaving far too much unsafe air, even on days with lower outdoor AQI.
Nerd note: In calculating the percentage of time of unclean air in the main text, I accidentally excluded the shortest test day (12/11). So I re-did the analysis with this data included, and the new result was that 85% of hours were unsafe. This oversight makes little difference, but I think itâs important to note for completeness.
 The Xiaomi Test Compared to Other Recent Tests in the Same Room
Iâve published dozens and dozens of days of test data using a Dylos particle counter in this exact same room (1, 2, 3, 4, 5, 6). That makes me pretty confident these poor results are not some basic flaw of the test design. However, there is always the possibility that the particle counter will break or lose its accuracy slowly over time. Or maybe some neighbor is setting up a secret chuanâr stand nearby. Fortunately, I have data from just two weeks earlier in the same room with the same particle counter.
If something is wrong with the particle counter, we should get weird results for these tests. But over several tests with a new-and-improved Original DIY 1.1, the results are pretty much what Iâd expectâa modest percentage higher than results for the DIY 1.0. Hereâs a sample test day for the DIY 1.1 two weeks before the Xiaomi test:
That test data looks normal to me. And that suggests that thereâs nothing strange going on with the room or the particle counter recently.
Converting Particle Counts to PM2.5 Micrograms
The tests used a Dylos Pro particle counter. To convert to PM2.5 micrograms, I took the 0.5 micron particles divided by 100. This formula comes from Dylos, and our prior tests show itâs pretty accurate compared to the US Embassyâa correlation of r = .90.
Original Data: Xiaomi
Below is the original data for the Xiaomi tests, including the outdoor PM2.5 levels as recorded by the US Embassy. My comparison data for the Blue Air, Philips, IQ Air, and DIY Cannon are available through in my earlier post. To request a copy of the original data in Excel format, send an email to [email protected].
 Original Data: Noise + Particle Count Test
Here is the original data for the test done in a different room with a different Dylos while decibel readings were taken:
Xiaomi Noise + Particle Count
Original Data: The Xiaomi Canât Stay on High?Â
Below are the transcripts from Xiaomi customer service where we asked whether or not the Xiaomi 2 can stay on high mode continuously.
Transcripts with Xiaomi Customer Service About Auto Mode
Original Data: Cannon with HEPA + Pre-Filter
As a part of the calculation of the percentage of hours with unsafe air, I included tests of the Cannon with a HEPA + pre-filter. When I ran these tests, the question I wanted to answer was whether adding a pre-filter decreases effectiveness (because pre-filters add to air resistance, with the upside of lengthening HEPA lifespan). The data showed only a minor decrease in performance: 2% lower average room effectiveness for 0.5 micron particles and 1% lower performance for 2.5 micron particles.
New DIY Design Increases CADR 15% - Supplemental data
Methods
Results were identical to my earlier tests of the Original DIY, Blue Air, IQ Air, Xiaomi, and Philips. Anna conducted the overnight tests in her 15m2 Beijing bedroom. Doors and windows were closed during the test (except for occasional opening during the first couple hours before bedâthat random variation is one reason I use the average of the last four hours to calculate effectiveness).
Anna used the Dylos Pro DC1700 to take a baseline measurement of the particulate in her room before turning on the purifier. She then turned the purifier on and left it on all night. The Dylos takes measurements every hour in hourly mode. I calculated effectiveness as (baseline particulate) â (the average of the last four hours before waking up).
 Daytime Tests
Itâs more convenient for Anna to run tests during the day, so she first ran 8 daytime tests. When I discovered this, I asked her to run nighttime tests instead because all of our previous tests were nighttime tests, and I didnât want to unnecessarily change a variable. So Anna then ran 11 nighttime tests.
My worrying seems to have been unnecessary. The average of all tests combined was 84% of 0.5 micron particles and 93% of 2.5 micron particles. The average of the night tests alone was 85% and 92%. These averages are different from the averages in the main article becauseâŠ
 Outdoor Pollution Fluctuations
I argued in the main article that room tests are a coarse method for testing for small differences in effectiveness between two machines. Random fluctuations in outdoor PM2.5 during the tests might add enough noise to the data to outweigh small differences in effectiveness. (But room tests do consistently reveal large differences in effectiveness. These room tests have routinely found differences in effectiveness between the Original DIY and the Cannon. Theyâve also spotted consistent poor performers like this DIY knockoff with a low-performing HEPA and the Xiaomiâs auto mode.)
So to detect small differences, I put more weight on the CADR tests. And for that reason, I tried extra hard in this data set to isolate days with stable outdoor PM2.5.
First, I pulled out all of the test days where outdoor PM2.5 fluctuated no more than 50 micro-grams during the test. That left 6 tests with an average outdoor PM2.5 of 90 micro-grams, which is right near Beijingâs average for the last few years. I report this result in the main article: 86% and 94%.
Another way to do this is to pull out the day with the most stable outdoor PM2.5. The data from January 16th stick out for being remarkably stable, with an average of 122 micro-grams. The results from that day are almost identical to the average in the main article: 86% and 95%.
But wait, itâs not a fair comparison if Iâm pulling out the very stable days for the DIY 1.1 and not the DIY 1.0. Indeed! I did the same analysis in my most recently published tests of the DIY 1.0 (tests of the slightly smaller 29mm HEPA). In that dataset, the results on the 8 stable days where outdoor PM2.5 fluctuated no more than 50 micrograms was 84% and 91%. Thatâs nearly identical to the 84% and 92% values for the DIY 1.0 here. Thus, Iâm moderately confident that the room tests are picking up on a difference in effectiveness between the 1.0 and the 1.1 (although I give the final say to the CADR tests).
 But Wait, Donât We Need a Control Condition?
How do we know particulate decreased because of the purifier? Maybe itâs because Anna was breathing in all that particulate. Maybe particulate goes down just because the doors and windows were closed. Maybe outdoor particulate went down during the test.
These are fantastic sources of skepticism. To test for these alternatives, Iâve run tests using a nearby control room, which rules out some of those alternative explanations. Some people suspect that PM2.5 is naturally lower at night (thatâs what I always thought), but PM2.5 data in Beijing and other cities shows that the opposite is true. PM2.5 is on average higher at night. That means running overnight tests actually makes it harder to show an effect (on average). But the better way to answer that question is to look at data from shorter tests and tests with the control room rule out the idea that these reductions are because of changes in outdoor air pollution.
Finally, are humans good purifiers? Iâve analyzed particulate data from rooms with and without humans to test this question. The results are abysmal. I donât recommend buying a human purifier.
 Average Outdoor PM2.5 During Tests
As Iâve argued before, itâs important to analyze the average outdoor PM2.5 during the tests. Thatâs because EVERY purifier Iâve tested achieves lower percentage reductions on days with high outdoor pollution. Itâs harder to fight particulate when there is really thick particulate seeping into the room.
For the 6 stable AQI days, the average outdoor PM2.5 was 90 micro-grams, which is right around the average in Beijing for the last few years. Over all 19 tests, the average outdoor PM2.5 was 121 micro-grams. These values are similar to values during tests with the Blue Air (113 micrograms) and the Cannon (111 micrograms).
 Original data: DIY 1.1
The attached ZIP file below is the original data for the DIY 1.1 tests for different days, including the outdoor PM2.5 levels as recorded by the US Embassy. My comparison data for the Blue Air, Philips, IQ Air, and DIY Cannon are available through in my earlier post. To request a copy of the original data in Excel format, send an email to [email protected].
DIY 1.1 Raw Data

Anya is live and ready to show you everything. Watch her strip, dance, and perform exclusive shows just for you. Interact in real-time and make your fantasies come true.
Free to watch âą No registration required âą HD streaming
SICKPAPES SPECIAL EDITION ON SUPPLEMENTAL DATA!!!
Supplemental Text 1 from:
Deshpande, G., Zhou, K., Wan, J.Y., Friedrich, J., Jourjine, N., Smith, D., Schedl, P., 2013. The hedgehog Pathway Gene shifted Functions together with the hmgcr-Dependent Isoprenoid Biosynthetic Pathway to Orchestrate Germ Cell Migration. PLoS Genet. 9, e1003720.
Like many people, I have conflicted views about the reams of Supplemental Data that accompany the online versions of most papers published these days. At its worst, Supplemental Data is nothing more than a grainy video of a bat performing fellatio, overdubbed with thumping techno.  At itâs best, however, a truly great Supplemental Figure reminds me of the famous scene in When Harry Met Sally where Meg Ryan loudly fakes an orgasm in a crowded restaurant, except that for me, the orgasm is real, and in those moments of ecstasy I am neither woman nor man but purely divine flesh of the GodBody incarnate. This Supplemental Text Document right here is one of those good ones which make me very glad to be living in the âEra of Supplemental Data,â as it was recently dubbed by Francis, the first pro-Open Access Pope.
To be real with you, Iâve never seen a Supplemental Document quite like this one. While Supplemental Data most often includes extra methodological details, additional controls, large datasets, or tangential experiments requested by reviewers, this one is a long-form prose essay about a decade-old scientific disagreement between these authors and a different lab. We here at SickPapes donât take sides in this quagmire - both of these labs are outrageously hot and time-tested - but we find it surprisingly compelling to read such an emotionally honest and open piece about a genuine scientific disagreement.Â
Our story begins in 2001, when Paul Schedlâs lab published a paper in Cell providing evidence that a secreted protein called hedgehog is involved in guiding embryonic germ cells as they migrate towards the future gonad. In one key set of experiments, they ectopically expressed hedgehog in abnormal locations, and showed that germ cells migrated incorrectly. This was interpreted to suggest that hedgehog is sufficient to influence germ cell migration.
As far as the public was concerned, the next thing that happened was In 2007, when Ruth Lehmannâs lab published a rebuttal, with the unambiguous title âhedgehog does not guide migrating Drosophila germ cells.â In the Lehmann lab, the hedgehog ectopic expression simply did not affect germ cell migration as it did in the hands of the Schedl lab members. This was not a matter of subtle differences in methods, this was literally a direct repeat of a simple experiment, giving different answers. Both labs are highly respected, and this discrepancy was hard to explain without getting a tad bit disrespectful.Â
Which brings us to explaining why this new Supplemental Essay from Schedlâs lab is so sick: it explains what was happening behind the scenes at Cell the whole time. First, they reveal that the Lehmann labâs 2007 paper actually originated as a technical comment sent to Cell in 2002 in response to the original 2001 article. The Cell editors asked the Schedl lab if they wanted to retract their paper, and the Schedl lab said âNo.â Instead, they came to an agreement with Cell that an outside, independent researcher would repeat the hedgehog experiments, communicating only with Cell and not with either of the labs in question (Shout out to Stephen DiNardo, stepping in to do a major solid for the scientific community without any personal glory).
After an extended drum-roll, Dr. DiNardo told Cell that his independent experiments confirmed the original results from the 2001 Schedl lab paper, not the Lehmann results. So, what does Cell do with this important finding? Nothing. Instead of publishing this informative back-and-forth, they didnât make any of this public, and just let everyone gossip for a decade. And, according to the Schedl lab in this Supplemental Text, this gossip-filled silence has âundermined [their] credibility in the scientific community, jeopardizing [their] careers.â Damn, Cell - that is some cold shit. I wish more folks would publish Supplemental Emotionally Honest Essays detailing the strife that various journals put them through. At the least, this might add fuel to Randy Schekmanâs dope protest against journals like Cell.
With that, we here at SickPapes wish to salute any and all Sick Supplemental Data, and wish you a very happy 2014!