In other news...

[Old Man]

But it's still awesome.  Seriously it looks like you put a lot of work into that.  If you can fix the normalization, label the size and color in the top graph, and maybe make the attributions more prominent, I'd give you a Nobel Prize.

 

I mean, the prize would be posthumous after your chart makes Facebook go supernova, but still.

[Sociotard]

Hmm, I'll look for a longer killed-by-police database. I could just trim them both down, but the police-killed database really suffers from sample size. For the smaller states, even one extra fatality can cause wild shifts.

 

So the graph needs a better legend. Or labels? I'm afraid of overcrowding the thing, so maybe not labels.

 

MLA format citation more prominent, can do.

 

 

[Ranxerox]

4 hours ago, Ternaugh said:

 

Your data isn't normalized (3 year period for killed by police, 10 year period for police being killed), so the graph isn't very useful. 

 

I disagree.  It is all normalized down to killing per year.  Yes, the data from a ten year period is going to be do a better job ironing out statistical outlier years than just looking at 3 years, but 3 years will usually be enough to give a person a sense of things.

 

So, yes, IMHO this chart and graph do a good job showing how dangerous each state is to it law enforcement officers and how dangerous those law enforcement officers are in turn.

[gewing]

On 3/25/2018 at 6:01 AM, Cygnia said:

Virus fished from pond cures man’s deadly antibiotic-resistant infection

Phages?

[gewing]

On 4/9/2018 at 2:26 AM, Badger said:

If he is a danger to himself, it quite likely they are unhinged enough to be a danger to anybody hapless enough to wander into that situation.  Best not to predict crazy.

 

In any case, the last fatal shooting in my area.  Some guy was breaking into a house, shot at the homeowner (who both called 911 and returned fire) and when the police arrived the suspect was in the process of leaving and started firing on the police.  SO, while I do try to take these things on a case by case basis, I will unapologetically lean towards the side of the police until convincingly proven guilty.  

Suicidal behavior is homicidal Behavior by default is what I was told in a training

[Cancer]

11 hours ago, Sociotard said:

Hmm, I'll look for a longer killed-by-police database. I could just trim them both down, but the police-killed database really suffers from sample size. For the smaller states, even one extra fatality can cause wild shifts.

 

So the graph needs a better legend. Or labels? I'm afraid of overcrowding the thing, so maybe not labels.

 

MAL format citation more prominent, can do.

 

 

 

I'll look at this once I'm at campus, not looking at it on a tiny-screen pad, and with more access to analytical tools and data sources.  

 

EDIT: Is "rate of gun ownership" guns owned (or sold) per unit population, or a fraction of households which have one or more guns?

[Sociotard]

13 minutes ago, Cancer said:

 

I'll look at this once I'm at campus, not looking at it on a tiny-screen pad, and with more access to analytical tools and data sources.  

 

Awesome. Let me know if you want raw data instead of a jpeg.

 

Does the table below need some indication of spread? Standard deviation or interquartile range or something?

 

I did find the database www.fatalencounters.org/ which goes over a much longer timeframe. I could cut that to match the ten years I have for police killed.  The only downside is that I'm not sure how credible it is. The Washington Post database has all the clout of a major newspaper, and this one . . . well it seems to have good citations.

 

I think I'll also switch the "gun control laws" database to one provided by the NRA. I mean, normally I'd just assume they'd just be a perfect inverse of a liberal politician (her A would be their F, etc) but I can still imagine conservative types being upset at seeing that citation.

[Sociotard]

25 minutes ago, Cancer said:

 

I'll look at this once I'm at campus, not looking at it on a tiny-screen pad, and with more access to analytical tools and data sources.  

 

EDIT: Is "rate of gun ownership" guns owned (or sold) per unit population, or a fraction of households which have one or more guns?

Fraction of households which have one or more guns.

 

I can edit the bar to make that more clear

[Michael Hopcroft]

OK. we're getting a bit sidetracked here, and it's all my fault.

[Old Man]

 But it’s a really good sidetrack. 

[Cancer]

Yeah, don't mind this one at all.  

 

It seems likely I won't have the whatzis to pursue this today, though.  Got to campus.  One office they are making the second attempt to replace my desktop machine (the first attempt had nothing I specified installed on it).  In the other office the new (last November) Mac seems to have died: when I try power-up the music chord plays, the "background" screen light comes on, and then ... nothing more.  Tried this a couple times (using power-off by holding the power button down).  Got someone coming in 2.25 hours to look at that box.

 

I'm camping for a little bit in our supply room/emergency adjunct office, which does have a generally unused working machine in it.

[Cygnia]

For Decades, Something Strange Was Stopping the Hearts of Gorillas in Captivity, The Cleveland Metroparks Zoo is Beginning to Trace the Culprit

[Starlord]

Cool.

[RDU Neil]

23 hours ago, Sociotard said:

I put together a little chart. I've posted a couple of times about how the reason the US has a higher rate of police killing people is because the police in the US are more likely to be killed themselves. Before I put this in my "ready to flamewar" pile, I thought I'd ask for constructive criticism?

Is the chart too busy? I have "rate of gun ownership" as the size of the bubbles, and "amount of gun control" as color. Maybe just "death vs. death" would work better.

Should the table below be sorted differently? Here it is grouped by degree of gun control, then sorted by state name. Should I do just by state name? or grouped by law type and sorted by Gun ownership rate? Sorted by one of the fatality kinds?

 

Any thoughts would be appreciated, is what I'm saying.

 

Am I reading this correctly, that minimal gun control and higher # of guns per household correlate strongly with increased number of both police deaths and deaths by police shooting?

 

Nice work on this all the way around, but just checking to see if I'm interpreting it correctly.

[Sociotard]

My principle argument was that Police killing people correlates with people killing Police. I made it in response to an article from Vox:

https://www.vox.com/2018/4/9/17205256/gun-violence-us-police-shootings

 

That article argued that Police shooting people correlates with number of guns and laxity of gun laws (and made no mention of people killing police). I felt that police dying would be a better indicator of police being in danger, so I made the graph you saw.  Now, there is some correlation with number of guns and gun law laxity, but the outliers are much more obvious.  Look at Ohio, As states go it has a very moderate concentration of guns, and Giffords gave its gun laws a D, not an F. Idaho has more gun-owning households and looser laws, but its death rate for police and civilians is far lower.

 

So yes, there is a correlation between gun ownership rate/ degree of gun control and fatal interactions between police and civilians, but it isn't as strong as the correlation between police killed and killed by police.

[RDU Neil]

24 minutes ago, Sociotard said:

but it isn't as strong as the correlation between police killed and killed by police.

 

Ok... I tried to follow the link you provided... https://ucr.fbi.gov/leoka

 

Following through, I'm assuming you are using the numbers of Officers Feloniously Killed information (taking out accidental deaths)?  https://ucr.fbi.gov/leoka/2016/officers-feloniously-killed/felonious_topic_page_-2016

 

 

[Sociotard]

Feloniously Killed, yes.

https://ucr.fbi.gov/leoka/2016/officers-feloniously-killed/tables/table-1.xls

[Cancer]

My Bad Tech Day (the Mac in my Physics office failed in the morning, and in the afternoon they took my old box in the Core Curriculum office and gave me  a Windows 10 machine in its place) utterly prevented me from doing anything today, including any acting on my promise of looking at this.  We'll see how tomorrow goes.

[Bazza]

 

[Sociotard]

Well, darn. I was going to use Fatalencounters.org for the "police killing people" database (it covered more years, including 2007 to 2016), but going through their descriptions, they lumped in all events where police were involved one way or another, not just those where an officer killed someone while performing their duties. The fatalencounters.org one includes a case where a lady was struck by a car fleeing police, a case where a cop killed his family and himself, and a case where a man committed suicide when confronted by police after a road rage incident. I don't think I'm sufficiently committed to this to go through several thousand events marking which ones count.

[Cancer]

On the Cancer's Bad Tech Day front, I seem to be able to do things again, though I had to submit to forced migrations from Firefox to Chrome and from python 2.7 to python 3.4.

[Old Man]

3 hours ago, Sociotard said:

Well, darn. I was going to use Fatalencounters.org for the "police killing people" database (it covered more years, including 2007 to 2016), but going through their descriptions, they lumped in all events where police were involved one way or another, not just those where an officer killed someone while performing their duties. The fatalencounters.org one includes a case where a lady was struck by a car fleeing police, a case where a cop killed his family and himself, and a case where a man committed suicide when confronted by police after a road rage incident. I don't think I'm sufficiently committed to this to go through several thousand events marking which ones count.

 

Surely there's another decent source for police shooting stats?

[Sociotard]

49 minutes ago, Old Man said:

 

Surely there's another decent source for police shooting stats?

 

Thus far I have found:

 

Washington Post: Solid reputation, has the kind of data I need, only goes back to 2015.

 

Fatal Encounters: No reputation but cites its work well. Data needs better sorted. goes back to 2000!

 

Justice Department: Kidding! I found articles explaining they are starting to collect data, but that means it won't go back far.

 

Several city and state scale databases: good, but too limited in scope. 

[Lord Liaden]

The American Center for Disease Control used to track those numbers. But since 1996 Congress has made that impracticable: https://www.theatlantic.com/health/archive/2018/02/gun-violence-public-health/553430/

[Cancer]

On 4/11/2018 at 7:26 AM, Sociotard said:

 

Awesome. Let me know if you want raw data instead of a jpeg.

 

Does the table below need some indication of spread? Standard deviation or interquartile range or something?

 

I did find the database www.fatalencounters.org/ which goes over a much longer timeframe. I could cut that to match the ten years I have for police killed.  The only downside is that I'm not sure how credible it is. The Washington Post database has all the clout of a major newspaper, and this one . . . well it seems to have good citations.

 

I think I'll also switch the "gun control laws" database to one provided by the NRA. I mean, normally I'd just assume they'd just be a perfect inverse of a liberal politician (her A would be their F, etc) but I can still imagine conservative types being upset at seeing that citation.

 

Found a trivial issue as soon as I could see it on a big screen.  The legend of the X-axis of the figure says per 1000; the heading on that column in the table says per 10,000.  Haven't had time to play more with the numbers.  But at least I am back up to two working computers I can use!