« first day (2059 days earlier)      last day (2726 days later) » 

06:28
So I was looking at how average asker-rep changes over time over the past 13 months and check this out:
$qes = EventSeries[{FromUnixTime@#["creation_date"], #["owner"][
        "reputation"]} & /@ Normal@$qs // Select[#[[2]] < 100 &]];
MeanFilter[
  TimeSeriesWindow[$qes, Today - QuantityArray[{12, .5}, "Months"]],
  10^6] // DateListPlot
We can pretty clearly see the start of the semesters, I think
Obviously everything is higher, the older it is, so I think it's only really meaningful to do comparisons within a few-month range
But the Sept / Oct and Jan / Feb dips are tantalizingly sharp
06:44
(Note that I selected asker rep < 100 to try to better pick up the effects of one-and-done type users, who dominate the user landscape in any case)
07:14
@b3m2a1 How about median?
 
3 hours later…
09:47
@Szabolcs I guess you will be interested:
1
Q: Framework behind Graph plots. DynamicNamespace and friends

KubaAs shown in DynamicLocation usage, we can use it to quickly create very nice functionality, that it to refer to graphics' primitives relative coordinates without calculations of boundary etc. As shown in the answer for that topic, except of DynamicLocation many related things appear: DynamicN...

any if anyone has anything to add, let me know.
10:44
How reasonable is this question from a complete beginner? (Who appears to have not done much research.) A good way to find out would be to try to accomplish the same task in other systems.
10:57
In Sage,
def fun(n):
    (res, tol) = numerical_integral((1/2 + 1/2 * erf(x/sqrt(x)))^n * exp(-x^2), 0, infinity)
    return res

list_plot([(i, fun(i)) for i in range(0,10)])
Probably not the best way, but it works, and shows that it is fairly easy to figure this out as a complete beginner. Especially with Mathematica's superior documentation.
@Szabolcs Certainly you'd expect someone with such a question to search to docs with something like "numerical integral" and "plot" -- maybe even "plot integer", which gets you to ListPlot, which has an example of plotting NIntegrate.
Right—and that method even avoids the evaluation order problem.
11:44
1
Q: Which external services are compatible with GeoServer?

SzabolcsUsing the GeoServer option, it is possible to load map tiles from external services. Which services are compatible with GeoServer? How can I find such services? Does GeoServer follow a standard that has a name that I can google for? The documentation shows a few examples: GeoGraphics[{Entity...

FormFunction[$Failed]
The documentation mentions the "XYZ" format for tile servers.
> FormFunction::nvldform: FormFunction expects a rule, a list of rules, an association or a FormObject as first argument.
Then why
FormFunction[sfdsf[aasd, 2, asd]]
does not complain?
XYZ seems to be a de facto standard, also called Slippy Map Tilenames en.wikipedia.org/wiki/Tiled_web_map#Defining_a_tiled_web_map
12:04
These are mostly good: basemap.nationalmap.gov Click one, then check out the URL of the StartTile or EndTile for a given zoom level.
USGS[map :
   "HydroCached" | "ImageryOnly" | "ImageryTopo" |
    "ShadedReliefOnly" | "Topo"] :=
 "https://basemap.nationalmap.gov/arcgis/rest/services/USGS" <> map <>
   "/MapServer/tile/`1`/`3`/`2`"
GeoGraphics[{Entity[
   "AdministrativeDivision", {"California", "UnitedStates"}]},
 GeoServer -> USGS["Topo"]]
There's this, but what is {s}? carto.com/location-data-services/basemaps
@Szabolcs style?
I think it's the server number. THe style is in the URL.
Putting in numbers from 1-4 works.
right
carto[style :
   "light_all" | "dark_all" | "light_nolabels" | "light_only_labels" |
     "dark_nolabels" | "dark_only_labels"] :=
 "https://cartodb-basemaps-1.global.ssl.fastly.net/" <> style <>
  "/`1`/`2`/`3`.png"
GeoGraphics[
 Entity["AdministrativeDivision", {"California", "UnitedStates"}],
 GeoServer -> carto["light_all"]]
stamenBase[style_, format_] :=
 "http://tile.stamen.com/" <> style <> "/`1`/`2`/`3`." <> format
stamen[style : "toner"] := stamenBase[style, "png"]
stamen[style : "watercolor" | "terrain"] := stamenBase[style, "jpg"]
GeoGraphics[
 Entity["AdministrativeDivision", {"California", "UnitedStates"}],
 GeoServer -> stamen["terrain"]]
@Kuba DynamicGeoGraphics[Entity["City", {"Krakow", "Malopolskie", "Poland"}], GeoServer -> stamen["toner"]]
12:25
@Szabolcs Looks nice, though I am in Berlin lately. At least air is better.
I wonder how we could use several layers from different servers
12:39
@Szabolcs That one needs an API key IIRC.
12:49
0
Q: How can I use multiple map layers from different tile servers?

SzabolcsWe can use GeoServer to use map tiles from an external service. There are many compatible services: Which external services are compatible with GeoServer? How can we use multiple layers, possibly loaded from multiple different servers, in a way that integrates well with the GeoGraphics funct...

13:11
How can I convert a GeoGraphics to Graphics? Show doesn't do it.
It seems that the first argument of GeoGraphics is a Graphics, so First is good.
@Szabolcs Yes. The tail inside GeoGraphics just contains meta information and if your are only interested in the graphics itself, you can ignore them.
14:06
posted on September 07, 2017 by Greg Hurst

In our continued efforts to make it easier for students to learn and understand math and science concepts, the Wolfram|Alpha team has been hard at work this summer expanding our step-by-step solutions. Since the school year is just beginning, we’re excited to announce some new features. Coverage We’re continuously working to expand our list of [...]

 
1 hour later…
15:10
@kirma this is what I got there:
MedianFilter[
  TimeSeriesWindow[$qes, Today - QuantityArray[{12, .5}, "Months"]],
  1*10^6] // DateListPlot
@Szabolcs I demoed Mapbox here in the chat a while back: chat.stackexchange.com/transcript/message/38192869#38192869
15:27
@b3m2a1 I assume this is again on reputation of 100 or less; how about all? Median might clean up data on that regard automagically - or not. ;)
15:50
1
Q: Making duplicate, but differently worded, questions more findable

b3m2a1I was tempted to open a question like "How can I make an arrow stick to a graphics primitive?" or something like that, then get it pushed as a duplicate to here: Framework behind Graph plots. DynamicNamespace and friends My reason for doing this would be to make it easier to find the Q/A if you w...

@kirma first off, since it seems you are interested by this too, here's my dataset: CloudObject["https://www.wolframcloud.com/objects/b3m2a1/mse_questions_8|7|16-9‌​|7|17.mx"]
Just CloudImport that
As for all the users, for some reason I still had to drop some to get MedianFilter to work. I dropped those with rep above 100000k because there are so few of them.
$qes = EventSeries[{FromUnixTime@#["creation_date"], #["owner"][
        "reputation"]} & /@ Normal@$qs // Select[#[[2]] < 100000 &]];
MedianFilter[
  TimeSeriesWindow[$qes, Today - QuantityArray[{12, .5}, "Months"]],
  Quantity[1, "Months"]
  ] // DateListPlot
So we can see a slow march-to-the-grave over the course of like Oct-to-Jan, but smaller variations get washed out. Not sure where that decline of those months is coming from.
Oh here's something kinda-fun:
$qes = EventSeries[{FromUnixTime@#["creation_date"], #["owner"][
        "reputation"]} & /@ Normal@$qs //
    Select[Between[#[[2]], {20000, 300000}] &]];
MedianFilter[
  TimeSeriesWindow[$qes, Today - QuantityArray[{12, .5}, "Months"]],
  Quantity[1, "Months"]
  ] // DateListPlot
We can see Szabolcs / Mr.Wizard / MichaelE2 were relatively more active last October and November (or some group of people in that over 100000k cohort)
 
3 hours later…
18:35
@Kuba should this really be tagged bugs? mathematica.stackexchange.com/q/155177/38205
I feel like the OP should wait to hear back from WRI before tagging it as such
 
1 hour later…
19:42
@b3m2a1 I think so, has he asked support?
I made a comment below the answer, let's see.
 
1 hour later…
20:47
Mike Foale's Machine Learning Flight System
http://community.wolfram.com/groups/-/m/t/1179035
 
1 hour later…
22:10
I have a large (3000 by 4000) integer packed array. I need to zero out those elements that would be white (or black) on a superimposed checkerboard. What's a fast way?
Can you improve on this without making it too complicated?
In[43]:= data = RandomInteger[2^14, {3000, 4000}]; // AbsoluteTiming

Out[43]= {0.165564, Null}

In[45]:= data UnitStep@
    Array[(-1)^(#1 + #2) &, Dimensions[data]]; // AbsoluteTiming

Out[45]= {0.546442, Null}
The bigger task is to de-mosaic Bayer-pattern raw data.
 
1 hour later…
23:35
@C.E. You used to have a tool for turning questions with the bugs tag into a nice formatted list but it was only a drop-box link that doesn't work anymore.
I couldn't find it on your GitHub account. Is this somewhere available?
@Szabolcs Factor 6 faster on my machine:
@halirutan The bug tracker is also available on packagedata.net/bugtracker/mma-bug-tracker.zip
data*Mod[Outer[Plus, Range[3000], Range[4000]] + 1,  2]; // AbsoluteTiming
@C.E. Ah, I saw this one but thought it is something different. Thanks.
23:59
@Szabolcs This is much faster:
m[[;; ;; 2, ;; ;; 2]] = 0;
m[[2 ;; ;; 2, 2 ;; ;; 2]] = 0;

« first day (2059 days earlier)      last day (2726 days later) »