In a comment on a prior thread (h/t CementAFriend), it was requested to use seasons shifted one month to better align with the experience of folks living in Australia. In this posting, I’m going to use those seasons. I’m also going to put up the 12 individual months along with the 4 seasonal averages (so 16 total graphs, all Australia & Pacific Islands).
Set Up
There were three “insane” temperatures found in the Region 5 data. Those have been set to “missing data flags” of -99.9 C so will not have any effect on these graphs.
Here’s the code to reset those values to missing data:
chiefio@PiM3Devuan2:~/SQL/bin$ cat FixAusBad.sql UPDATE temps3 SET deg_C=-99.9 WHERE stnID='50291652000' AND month=" APR" AND year=2000 ; UPDATE temps3 SET deg_C=-99.9 WHERE stnID='50396237000' AND month=" AUG" AND year=1992 ; UPDATE temps3 SET deg_C=-99.9 WHERE stnID='50998644000' AND month=" MAR" AND year=2012 ; [...] MariaDB [temps]> source FixAusBad.sql Query OK, 1 row affected (0.72 sec) Rows matched: 1 Changed: 1 Warnings: 0 Query OK, 1 row affected (0.03 sec) Rows matched: 1 Changed: 1 Warnings: 0 Query OK, 1 row affected (0.04 sec) Rows matched: 1 Changed: 1 Warnings: 0 MariaDB [temps]> [...] MariaDB [temps]> SELECT deg_C FROM temps3 WHERE deg_C >50 AND region=5; Empty set (36.28 sec) MariaDB [temps]>
Then, with that fixed, I’ve re-run the anomaly creation step and we’re ready to make the monthly and seasonal graphs.
MariaDB [temps]> DROP TABLE anom3; Query OK, 0 rows affected (1.09 sec) MariaDB [temps]> source tables/anom3 Query OK, 0 rows affected (0.31 sec)
(The table layouts and load anomaly code are documented in earlier postings)
The Seasons
The seasonal mapping used in this set is:
Summer January, February, March Fall April, May, June Winter July, August, September Spring October, November, December
So first off, the seasonal graphs. I’ve put the word “Local” or the letter L in the names of these to designate that the seasons are divided as the locals do it. All of the graphs have the same range of -6 to +4 anomaly as this lets the graphs be compared more easily. It compresses some of the lower volatility months – or perhaps just better demonstrates the lack of volatility… Especially compared with some of the continents in the prior posting.
Local Summer Australia Pacific Anomaly GHCN v3.3
Local Fall Australia Pacific Anomaly GHCN v3.3
Local Winter Australia Pacific Anomaly GHCN v3.3
Local Spring Australia Pacific Anomaly GHCN v3.3
The Months
I’m not going to put much commentary here as I already commented in the earlier posting. So, without further ado, here’s the months:
January Australia Pacific Anomaly GHCN v3.3
February Australia Pacific Anomaly GHCN v3.3
March Australia Pacific Anomaly GHCN v3.3
April Australia Pacific Anomaly GHCN v3.3
May Australia Pacific Anomaly GHCN v3.3
June Australia Pacific Anomaly GHCN v3.3
July Australia Pacific Anomaly GHCN v3.3
August Australia Pacific Anomaly GHCN v3.3
September Australia Pacific Anomaly GHCN v3.3
October Australia Pacific Anomaly GHCN v3.3
November Australia Pacific Anomaly GHCN v3.3
December Australia Pacific Anomaly GHCN v3.3
So there you have it. Australia and the Pacific Islands all sorts of detail. To me it looks like maybe a bit of UHI in winter but mostly like nothing much happening, and summers even showing a bit of a cooling over the years. Especially January and March.
UPDATE – Airports vs NOT Airports:
I made a couple of graphs of Airports vs NOT Airports for Summer in Australia / Pacific (remember that Region 5 includes places like New Zealand, the Philippines and other islands too) Here they are for whatever they mean. Both look to have some rise at the end, but the NON-airports do it by narrowing the range as their number drops a lot (see comments below) and their elevation drops (see comment below). Airports seem to have a more defined “rounding up” rise with the growth of air travel and the Jet Age.
NOT Airports Summer In Australia Pacific Anomaly GHCN v3.3
Airports Summer In Australia Pacific Anomaly GHCN v3.3
Update 2 – Elevation Over Time:
Here is a graph of station elevation over time for Region 5 – Australia Pacific.
Station Elevation Average over Years for Region 5 – Australia Pacific Islands
Musings from the Chiefio 
Thanks EM. Great work. I wish I had your computer skills. My last computer skill was with Basic on an Apple 11 when I wrote a speadsheet program with four (or more) dimensions before Microsoft developed Visicalc.
Great to see that there is no temperature increase in Australia.
It appears to me that an odd scrunching of anomaly (significant reduction in scatter) in the January scatter plot in the early 2000’s (about last 10-12 years of the plot) like something changed on instrumentation about then.
Suddenly all the anomalies are positive and scatter over a very small range compared to other months in the same years.
E M ! Wow ! And thanks !
This is what I would love to see publically available from BOM here in Oz. But they do not make such date or analyses available. Only the the ‘scientific’ cultic elite have that !
Larry you are right about the plot for January. But every month, except May, August & October, shows that ‘scrunching’ over time ! July is particularly pronounced.
This could partially be the impact of ‘Urban Heat Island Effect’ as it does tend to reduce the minimum temperatures in urban locations. But how many of these weather stations are urban as opposed to rural ? I have no idea from the station numbers. So perhaps a link to the list of weather stations by name would help.
A separate line of thought re seasons : I’m especially glad you did this by month. Australia is huge and stretches from 40 degrees South to near the equator. So the concept of a unified set of ‘seasons’ does not make sense for the continent. So we all know that in the North there are two seasons ; the wet & the dry. But even in the South, at the regional level there is disagreement about seasons with some folks drawing input ( at the regional level ) from the Aboriginal lore on seasons. An interesting book on this issue is Tim J Entwisle’s “Sprinter of Sprummer” published in 2014 by CSIRO. He suggests that in the South Eastern region of Australia, there are 5 seasons : Winter in June/July; Sprinter in August/September; Sprummer in October/November; Summer from December to March; and Autumn in April/May.
Here in South Australia this view is reflected in the old farming view of the growing season : It starts with the ‘Break” ( the rains ) in Autumn which can happen any time from late March to early June. It ends with harvest in November/December – depending again on whether the growing season has been a wet one or a dry one. Last year 2018, was a dry one with crops ( wheat.barley, canola, etc ) harvested early or not at all due to drought. But 2016 was a wet growing season with bumper crops still being harvested in December/early January in places.
And as we all know local temperatures are hugely effect by the moisture in the air = rain !
Lots of moisture in the air stabilises temperatures. Dry air conditions promote temperature variations both high & low. BOM is talking about how hot 2018 was. But I had a huge number of cold frosty nights with low low temperatures. I lost lots of garden plants due to those frosts. So I know it was cold.
My home heating bill told me the same story. So BOM is spinning us a tall tale infested with bull dust in my EXPERIENCE..
Thx E.M. Oz BOM eat yr heart out. )
Will post this link @ Jo Nova.
Well one possible reason for the low going excursions to be reduced might be that the “high cold places” get reduced:
The Code:
As this whole table is 178 lines (years) long, I’m going to chop out some basically repetitive bits and leave in representative chunks.
So a drop of average elevation to about 1/2 of the 200 m in the 1930s. I’d also wager that 109 m elevation is closer to the ocean…
I tried to use the “proximity to water” field but it has negative values in the average as does the proximity to Urban – I think it’s another -999 missing data flag issue. I’ll work on it some more and get back…
Yup, there’s “-9” in my inventory file and chasing it back up stream it is in the original input data:
501 is Australia IIRC and if you scroll off to the right, it has -9 in a couple of places where there ought to be distance to ocean and distance to urban area. It doesn’t say that’s missing data in the README file, but given they use it in other fields, I’m pretty sure that’s what it means.
The important thing is it is not “my bad”, it is in the original archived inventory file. So I just need to figure out how to deal with it…
Or maybe it’s that airport thing…
Again as these are 161 rows long, some pruning of repetition will be applied: Not that a station that started life in 1855 as a grass field and became an airport late in life will still be marked as an Airport in 1855 since the inventory data does not know about years, only current condition. So this first report is showing just one thermometer and it is an Airport recently – 1855 not so much despite what the inventory file says..
So 108 Airports today to 62 non-airports today or 1.74 airports / rural in 2015.
And 203 airports in 1971 to 504 non-airports or 0.4 airports / rural in 1971.
I’d bet that matters…
Then there was that great dying of thermometers… it took out a lot of the non-airstations too:
EH, you’ve nailed it I think.
When you wrote ” So 108 Airports today to 62 non-airports today or 1.74 airports / rural in 2015.
And 203 airports in 1971 to 504 non-airports or 0.4 airports / rural in 1971.” I knew you had cracked it…..And then you followed up with the list of the great thermometer dying !
Meanwhile the Bureau of Misinformation has created it’s ACORN Sats 1 & 2. In both of them BOM makes the assumption that it’s current temperature stations are all correct and then proceeds to lower the temperature readings in the past.
In fact it should be doing exactly the opposite : Distrusting it’s current station readings and adjusting them inline with the correct station readings of the past.
I’m nominating you the Great Scientific Australian Elephant Stamp for 2019 !
EM not sure what you are saying about the table of named stations starting with Kalumburu. They are all in the north of Australia in a climate area of Tropical wet and dry or monsoonal. Further they all appear to be airports. Thursday Island is north of the tip of Cape York and not far from PNG. It has an Airport and is the centre of the Torres Straight Islands. Kalumburu and Troughton Island are in the north of WA both have airports. Except for Darwin and Katherine there are few people living there. In Darwin and Katherine November is the most unpleasant month because it is hot and dry before the monsoon rains. December can be variable with lightning starting bushfires. Gove and Weipa are mining towns on the coast with airports just about the only access (apart from ships)
@Cementafriend:
I’m not saying anything about them in particular. That is just a sample of the raw input data.
I was worried as I got negative averages for distance to the ocean or distance to a town when trying to find how the average changed over time. Then found -9 in the field frequently. Had I screwed up the data load? So I looked at the input file from GHCN and also found the -9 values.
Scroll the posted bit to the right and you find:
It is the “-9” values in that sample I’m pointing out, and claiming it looks like another “missing data flag”.
As the average is negative for all of Australia / Pacific it looks like the data on distance to water and distance to urban center are crap and not suited to use. One would need to go through the inventory of stations and figure it out by looking up each station.
Basically, I can’t use this data to plot how “average distance to ocean” or “average distance to urban center” changes over time.
BTW, in case I didn’t point it out, I’ve added an update at the bottom of the posting with a graph of “without airports” and one of “airports only”.
They are clearly different and that alone means that changing the airport/non-airport ratio will change the results, but it also looks to me like the non-airports (with reducing numbers and reducing average altitude) have a loss of low going anomaly (while the tops don’t get hotter) and the airports DO get hotter over the period of the Jet Age – i.e. the large growth of commercial aviation and burning tons of kerosene in take-off rolls and adding hectares of asphalt parking and. all the rest… They also have a very narrow spread as though asphalt in the sun was almost a constant ;-)
I’ve also added a graph of stn_elev over time showing a drop in average elevation right in that last up-turn of “warming” (or loss of cold excursions in non-airports).
The code to produce it is:
plt.title("GHCN v3.3 Australia Pacific Elevation") plt.ylabel("All Months Elevation Average") plt.xlabel("Year") #plt.ylim(-6,4) #plt.ylim(1850,2020) try: db=MySQLdb.connect(user="chiefio",password="LetMeIn!",database='temps') cursor=db.cursor() sql="SELECT year, AVG(I.stn_elev) FROM anom3 AS A INNER JOIN invent3 AS I ON I.stnID=A.stnID WHERE I.region=5 AND I.stn_elev>-9 GROUP BY year;" print("stuffed SQL statement") cursor.execute(sql) print("Executed SQL") stn=cursor.fetchall() # print(stn) data = np.array(list(stn)) print("Got data") xs = data.transpose()[0] # or xs = data.T[0] or xs = data[:,0] ys = data.transpose()[1] print("after the transpose") plt.scatter(xs,ys,s=2,color='purple',alpha=1) plt.show() except: print ("This is the exception branch") finally: print ("All Done") if db: db.close()One way to fix the season issue would be to align the calendar with the astronomical season start. So, for where I am, Winter would be Jan, Feb, Mar; with Jan 1st = to the old Dec 21st. Winter being about 90 days, all three months would have 30 days in them. That would take some getting used to, though. A catch would be autumn (89 days) and summer (95 days). Another catch would be leap years and which season gets the added day.
@CDQuarles:
Then there’s the minor problem that the data are already monthly averages so you can’t get into the individual days of the month…
One would need to drop back to the raw daily data that is not available easily globally…
E.M.: “[Airports] also have a very narrow spread as though asphalt in the sun was almost a constant ;-) “
Absolutely. Let’s say there is one day in a year where it is cloudless every year on that same day. A blacktopped area will give pretty much the same reading regardless of the temperature of the airmass over the area that day. I would think that the air mass would have to have an extreme temperature deviation, cold or warm, from the average temperature on that day to make much difference at all in the reading for that area. Otherwise, for a given day, the temperature above the tarmac is determined more by cloud cover.
BTW, is jet exhaust considered to be an extreme air temperature deviation? ;o)
Not for a runway, I hope. Conditions at the runway are crucial for safe flights. One should expect jet exhaust on/near a runway, I’d think. That, of course, means that weather reports from an airport’s runway should not be considered ‘representative’ of the environment away from the airport. ;p
@cd – I get your point about air temp for takeoff and landing. I was just kind of idealizing the situation. The sun is essentially constant, so if a surface doesn’t change year over year and there are no clouds on the same day year over year, then energy in and energy radiated out will be the same year over year on that day in that place. The only variable is ambient air temperature and wind over the area.
It’s really pronounced in winter on a cloudless, windless day on or near blacktop. It may be rather cold, but any snow on or around that blacktop disappears in a hurry.
I use that to my advantage to save on shoveling snow. If an inch or two of snow falls overnight and the following day is sunny – warmer or colder, it doesn’t matter at all – I’ll just make a few random paths up and down the driveway with a snow shovel and the rest will be gone by afternoon.
E M Thanks for including the charts UpDate to this post, re elevation and airports versus non airports.
Elevation clearly has an impact. And reducing the number of weather stations in the set from higher elevations does bias average temperature upwards.
Clearly whether a weather station is an airport or not also matters. Aircraft runways and engine exhausts and probably climate control inside the terminal buildings has an impact. ( Cooling a terminal means pumping a huge amount of hot air out of the A/C units mounted on the roofs. But I wonder, does warming the interior of terminal buildings with A/C cool things down outside ? )
Interior heating is usually NOT from a heat pump but from burning natural gas or oil or using electricity all of which make “waste heat”. Airports are hot. Period.
Not here in Oz in my experience EM. The terminals I know here use reverse cycle A/C.in Winter.. Warmer Winter temperatures maybe with hardly any times below 0 degrees C. And gas & oil are too expensive – or have been in the past compared to to the cost of electricity…from coal !
Valentia Observatory in Ireland is regarded as one of the best-placed stations in the world, and has reliable long-term records.
Have a look at their Unadjusted data covering nearly 130 years.
GISS Surface Temperature Analysis
Station Data: Valentia Obse (51.93N, 10.25W) ID:621039530000
https://data.giss.nasa.gov/cgi-bin/gistemp/stdata_show.cgi?id=621039530000&dt=1&ds=5
The unadjusted chart for Valentia, says it all. The ‘homogenised’ shows the extent of the lying.
It’s worse, if memory serves me correctly Bill in Oz. I believe the 1880 start selected for the Valentia data shown was for the same reason suggested by Chiefio at the following link.
@hillbilly That chart ends at 2000. I wonder what it would show now in 2019 ?
Ever rising off the chart ?
Or a peak and then a decline as per the sine wave you suggest ?
Geeeze who ever did that chart is nuts – I can only see one trace (black). I can just barely tell there is another trace there for the unadjusted but pale yellow (or green) on white is a completely idiotic choice of colors, the other two – ?? are they even on this chart?
I cannot even see the other two pastel colors on the chart, can anyone else see them?
@ Larry & Bill of Oz. Click on each category which highlights the little Line Colour bar. Then click on the lower line and continue up. Each click highlights the trace above it,
It spans 01/1880 – 02/2019.
In an article on his site many years ago, “What’s Wrong With The Surface Record”, the late great John Daly recorded Valentia Observatory records started in 1869. His work and words still resonate, particularly in regard to the necessity of using ‘greenfield’ weather station sites.This why Valentia was so ideal.
I note that the first “great Dying of Thermometers” occurred in around 1992 after temperatures had plummeted around the world. despite Hansen’s dire predictions in 1988.
http://www.john-daly.com/surftemp.htm
This NASA link shows the raw ( unadjusted ) temps in pale green and the homogenise data in black. ( There are 2 other colours that are irregular in functioning..And I do not know what they are measuring ) But they both go from 1720 to 2020…
https://data.giss.nasa.gov/cgi-bin/gistemp/stdata_show.cgi?id=621039530000&dt=1&ds=5
Yes Bill that is the image I was talking about, I literally cannot see that green against a white background, it is right at the limit of my perception I can “kinda” perceive something is there but the pale green is for all intents and purposes invisible to me (I am red green colorblind).
Just about the dumbest color combo they could pick. But web designers and especially scientific geeks are completely oblivious to good graphic design rules they routinely use color choices which are nearly impossible for me to see, or although I can see them I cannot distinguish between pairs of the colors. Like 3 traces one dark green one dark brown and one dark gray – if the lines cross I have no idea if they just kissed and then separated with or if they crossed and the one that started out on the bottom is now on the top or in the middle.
Sometimes I can fix it with a graphics program and force color changes on the back grounds or lines but sometimes even that does not work.
Or my most serious pet peeve is the actual color of the lines is not “exactly” the same color triplet as the key colors.
Sorry 1880 to 2020 !
But EM’s chart from 2009 link goes back to 1720… I wonder where EM got the date for the earlier years from ?
Just read through the comments..it came from here :http://www.smhi.se/content/1/c6/02/50/31/attatchments/upps_www.pdf
But that link has been cut..No longer available..
Ummmmm ? SMHI, Swedish Weather Bureau getting rid of the evidence ?
@Larry I appreciate your colour blind problem…The light green for raw data is a ad choice of colour..
Though so too are the other colours which seem to do nothing I can make sense of.
Just to set you at ease.re ” if the lines cross I have no idea if they just kissed and then separated with or if they crossed and the one that started out on the bottom is now on the top or in the middle.”
That do happen at all. In each & every year of the plot, the green line is always above the black line.
And that implies that NASA have also lowered recent year temperatures on the black homogenised line !
Dohhhh !
I figured out how to get enough color saturation to see the plot (before we were tipped off to an undocumented feature of clicking on random things highlighted colors)
Why do web designers think people have nothing better to do than mouse over every possible bit of info on the screen or click on it to see if it does something? We have a wonderful method of communicating that sort of information called a text note that tells you something exists – or maybe make the clickable thing actually appear to be a clickable thing.
Crappy web design just drives me nuts. I spend about 10 seconds looking at an illustration if it does not provide the info I am looking for I am gone, I am not going to beta test your web page to see if I can figure out how it works.
@Bill in Oz:
The folks at GISS decided that there were too few thermometer records prior to 1880 to be usable for a Global Analysis, so toss anything prior to that. Hadley use 1850 IIRC. There’s a fair amount of earlier data, but not global. You can see that in the graph of thermometers over chunks of time:
I probably ought to make one each that ends exactly in 1850 and 1880, but there’s one there that ends in 1901 and even it is too short of global coverage.
So the further back in time you look, the crappier your results.
@Color Choices:
Unfortunately, there often are not a lot of choices for colors. In Python there are only a few color words that work. You can get other colors via some kinds of numbers but I’ve not got that particular skill yet.
Larry: Are the “seasonal” graphs where I used green for spring visible for you? If not, I can redo them in some other color. (I tried purple on a few and liked it, but was drawn to green as symbolic of spring and less reading required for folks to figure out the seasons). Personally I find the yellow dots harder to see (so used “orange” for fall so it would be visible more easily and “orange” looked more yellow to me ;-)
I’ve figured out that these color words work: red, green, blue, yellow, orange, purple
Others, like magenta, do not. I’ve not found a definitive list (but stopped looking as soon as I had “enough” colors) nor have I twigged to whatever bit of hidden magic sauce is used to set color-by-numbers… (again, as I had ‘enough colors”). I found many examples that did really trick things like random colors and spectrum by height on an axis but nothing that was just “c=Saturation,Hue” or similar; so just lept to a few color words.
So if anyone knows how to set things like “DARK green” or “navy blue” or whatever other color in a python graph, just let me know…
Yes I can see the April, May and June plots above just fine the green is fully saturated and significantly darker than the background.
[https://chiefio.files.wordpress.com/2019/03/l6aupaanom.png]
I need more contrast than people with normal vision, I can see the colors but not as strongly as normal vision would. Think in terms of gray tone, the colors need to differ by more than 10% gray tone for me to reliably see them or discriminate between similar colors. That stupid light green on the plot we have been talking about would be such a light gray tone that it renders as the same gray tone as the background when you try to flood fill the background color on the plot image.
16 bit image color choices so not leave you much option so you just do the best you can, but your plots are no problem because you are not trying to get cute.
Color pairs that cause problems for me (an many red green color blind) are:
Easily confused or cannot be discriminated between:
dark brown and dark green
dark green and dark red ( I can walk right past a dark green pine tree with red survey flagging tape tied to a branch and never notice the red flagging tape – actual incident in search and rescue)
dark magenta and dark brown
dark blue and dark purple (same for pale purple and sky blue of similar shade)
dark red and black (web pages that use a dark blood red text to highlight a word – I will not even notice the red text but will just see a page of text I think is all black text)
light green and yellow
very pale pastel cyan on white background (I am almost totally blind to pure cyan)
Any combination of brown and green where both colors have almost the same gray tone
Any combination of red and black/gray where both colors have almost the same gray tone.
Any pair of light pastel shades of blue, turquoise, cyan, blue green, or pale purple of similar gray tone.
Etc.
For example a new international orange traffic cone is reasonably visible to me, but an old used dirty orange traffic cone with a reddish dirt background like an embankment I might not notice until I am right on top of it.
Flashing warning lights on emergency vehicles at long distance I cannot tell if a vehicles flashing lights are red or a dark orange amber. (ie is it a police car, or a tow truck), only when I get close and the image size is large enough to illuminate a large number of cones in the eye can I tell the difference (most of the time), some emergency amber lights are so close to some flashing red warning lights that I can never tell which is red or amber and have to use other visual clues to determine if it is a cop, firetruck ambulance or tow truck or utility service vehicle.
Red green colorblindness is caused by different sensitivity to colors by the cones in the eye, and possibly different numbers of cones (ie fewer cones sensitive to a given color)
https://enchroma.com/blogs/beyond-color/how-color-blind-see
https://enchroma.com/pages/types-of-color-blindness
I have simularities to both the above discriptions of Protan Color Blindness, and Deutan Color Blindness
I just took their online colorblind test and register as YOUR RESULT IS STRONG DEUTAN
https://enchroma.com/pages/color-blindness-test#&ui-state=dialog
WHAT IS DEUTAN COLOR BLINDNESS?
Deutan color blindness (also known as deuteranomaly) is a type of red-green color blindness in which the green cones in the eye detect too much red light and not enough green light.
As a result red, yellow, green, and brown can appear similar, especially in low light. It may also be difficult to tell the difference between blues and purples, or pinks and grays.
On color triplets a tool I use to translate colors into the color triplets (so I can tell which trace on a graph corresponds to what point on a key reference is called:
whatcolor4 (it allows you to translate a color on screen to either hex or decimal color triplets)
Unfortunately for you it is only for windows.
http://www.hikarun.com/e/
http://www.hikarun.com/e/wcol500e.exe
I assume python would have a feature to specify colors by the color triplets.
A quick look around I found this which looks like it might be useful for your needs.
View at Medium.com
A discussion of colors in web page design and good practice with colors
https://www.smashingmagazine.com/2016/06/improving-color-accessibility-for-color-blind-users/
@Larry:
From that “Practical Introduction” link:
“The matplotlib package has excellent support for managing colors and plots, however, it can sometimes be a bit tricky to find out how to combine the various different classes/functions in this package. ”
That’s an understatement and a half…
In their first examples they use a half dozen to a dozen various functions (Objects) and options. None is in the basic definition of the language. None I’d seen before. They are all in glue on “libraries” of stuff. Then there’s dozens of such libraries… That, IMHO, is the bane of all OO programming. You can learn the language, but then get to reinvent everything from scratch, or spend months more digging around to find out what’s commonly used in various libraries of objects, how to call them, etc. etc. A never ending decent into the depths…
Yeah, it is all great and wonderful if you have programmed a given problem domain for the last decade in that language and know where all the musty corners are with the goodies; but noobie hostile…
So thanks for the pointer, it looks very useful. (My only complaint so far is it leaps right into making log proportional color maps on one shade axis… doesn’t say much about “How do I make my ONE line magenta and show that in the legend?”…) I did find the example of “default line colors” in a line plot useful “out the gate”. I suppose I ought to have lept to the conclusion there were defaults…. but I Didn’t, I was planing on making line charts but stressing over no way to set the colors differently.
Now I have a shot at it…
It is kind of like getting a book and finding out it is mostly just an overview of several other books that are not included in the set. So no character development, no history, no cast of characters. Just suddenly Binny Frobush shows up in Act 3 and everyone suddenly runs away. WT? You were supposed to just know that in book 4 he had Plague… and is a carrier… Only worse, you are assigned the job of writing Act 4, the Cure, and don’t know that Magdalen Grimwald was given magical curing powers if given marigolds and SPAM in book 6 and if you just chant her name backwards she appears and cures everyone…
Oh Well. It works, kinda. Given a day or so per “trick” (method? function? Incantation?) learned and hopefully only needing a few dozen “tricks” for my graph drawing, I guess it is OK…
I haven’t read the whole description of color blindness effects above (but will). Just thought I’d point out that color sensed by a cone depends on cone length. Folks who see 3 primary colors have 3 lengths of cones to match the wavelength. Those with only 2 lengths see only the two colors (for any given 2).
Many birds have 4 lengths and see a very different color pallet. Some people also have 4, though more rare than colorblindness as it is easier to lose a bit of DNA than invent a new one.
Dogs, IIRC have 2 lengths that let them see blue, green in a grey with shades kind of way. Knowing about the cone length / color connection now lets us examine the eyes of critters and state just what they see… There are also some birds with only two cone lengths.
FWIW I get a sense of the “problem” every time I edit a file on Devuan. The “default” is to use some kind of color interpreter in vi (really vim but aliased) and it uses a very deep blue for “comments” in code. One that I can see is there, just can’t read on a black background…
I’ve not spent time to figure out how to change those color choices either. Limited time, unlimited “things to fix”… So when necessary, I’ll adjust the background color in lxterminal to read the comments (usually a brownish pumpkin color works, but then I can’t read other colors they chose for other line types). Once again someone being “helpful” adding color by default and causing things to break in a different context..
Oh Well…
Yes I “HATE” the default color option on linux, several of them are horrible, the dark red on a black screen background does the same to me, I can tell there is text there but I cannot read it.
First thing I do when I log onto a new box is run these unalias commands to eliminate that problem on the commands I use most that have color set to on
unalias vi
unalias ls
unalias grep
On our Centos systems these are the default aliases that set color to ‘auto’
alias egrep=’egrep –color=auto’
alias fgrep=’fgrep –color=auto’
alias grep=’grep –color=auto’
alias l.=’ls -d .* –color=auto’
alias ll=’ls -l –color=auto’
alias ls=’ls –color=auto’
You can change the color choices but it is a royal pain in the butt and I finally figured out the quick and dirty solution was just to unalias all commands that are aliased to –color=auto.
Looks like “ls” is aliased, but not vi / vim so a bit more to undo it…
Now vi is a heck of a trail that leads nowhere:
chasing through the man pages has a half dozen (or more) places where a config or rc file is set… So much for the KISS principle…
Looks like it has it’s own way…
https://www.cyberciti.biz/faq/turn-on-or-off-color-syntax-highlighting-in-vi-or-vim/
So the “magic sauce” is :syntax off (why in heck would ‘syntax’ refer to color…) and you must know that to put it into a .vimrc file to make that stick… What a horrible process choice.
Pingback: GHCN v3.3 vs v4 – Top Level Entry Point | Musings from the Chiefio