I’ve been working on my mapping skills lately in preparation for the first FlowingPrints poster, so when I came across this dataset for abortion rates in America, I had to map it.
The darker the shade of green, the higher the number of reported abortions per 1,000 live births.
New York has the highest rate with a whopping 507, which is a little over a third. That I’m not so sure about though. I’m thinking that there might be some high numbers in the ’70s driving that rate up, but I’d have to look deeper into that. Wyoming, on the other hand, only had a reported 14 abortions between 1970 and 2005.
In retrospect, the choice of green probably wasn’t the best color choice, but seeing as this is just practice, I don’t think it’s a big deal.
How I Made It
In case you’re wondering, I made the basemap in R using the maps and maptools packages. It was actually only 5 or 6 lines of code after I got the data how I wanted it. Then as I always do, I brought the PDF into Adobe Illustrator for some touch-ups and annotation.
Check out the full version here.
UPDATE: I revised the map using the Albers projection, so it doesn’t look so funky. Of course, it was more difficult than originally thought. Tutorial to come.
“New York has the highest rate with a whopping 507, which is more than half of births.”
Wouldn’t that be a little over 1/3? It’s 507 per 1000 live births. I believe that by definition, the two are mutually exclusive.
you’re right. fixed. thanks
It’s somewhat shocking to see abortion numbers like this, especially when juxtaposed with Gallup’s May 2009 polling data , which shows that for the first-time ever, a majority of American’s call themselves “pro-life.”
There’s got to be some cognitive dissonance going on with a large part of the electorate. It’s not hard to infer that many people who voice opposition to the procedure have either themselves had — or helped a partner, sibling, or close friend who had — an abortion.
Nate, I’d be interested to see how Gallup’s numbers — in terms of state by state support for abortion — correlate with the rates you’ve shown here.
Interesting. I’m curious about how to interpret the data: it may be dangerous to assume that the abortion rate among Wyoming residents is so low: could they all be heading to New York for their abortions? Are we seeing states’ tolerances to abortion, ability (through concentration of population, for example) to support abortion facilities, incidence of out-of-wedlock/young pregnancies? Would we expect the Utah rate to drop significantly now, given its apparent predilection for shooting doctors?
Projection! What is that unshapely United States?
But then again, I wouldn’t know where to start to build my own basemap, so I’m impressed. Just curious, what’s the advantage of building your own rather than downloading one that’s projected?
I think this is a good start.
Green might be a good color choice; if you’d used red or blue it would seem more politically charged (IMHO).
Also, is there no data for CA or LA? Those seem like big omissions to me and leave your narrative comments about NY & FL open to much debate.
Finally, 1970-2005 is a very wide range. To tell a more complete story I think it would help to offer comparative maps (maybe by decade, although that would throw off the 2000-2005 period).
Does reported mean “abortions performed in the state” or “abortions performed by residents of the state”? I’m just curious whether these numbers are taking into account people who have to travel out-of-state. (And, being from Wyoming, I would think that very likely the case.)
Is there a possible misrepresentation in this way of stating the numbers — i.e., say a state has a relatively low birth rate but is somehow an “easy” place to have an abortion (maybe people from several surrounding states go there). I think I would like to see a companion map that shows simply the birth rate for each state — maybe live births measured against total state population?
I think that’s an excellent point. Since it’s a per 1,000 base, it’s all still relative, but having that companion count at the same time may be somewhat revealing.
See this post on fivethirtyeight.com for a great discussion on the issues of mapping abortion rates by state:
Unfortunately, the data are not nearly as cut and dried as we would like them to be.
First, there are reporting issues…the number of abortions performed in your state is a politically sensitive matter. Then, there is the issue mentioned above by another Alex, that of residence. See the post linked above for all the details.
I think this would be neat to see in small multiples, to see the change over time.
Can you share your code and a snippet of the data in the format that worked?
i’ll post a tutorial once i get my junk in order
Ouch – you’re hurting my eyes! You must use a different map projection. Doing so will not change the values you want to map – but let you represent the sizes and shapes of your states with much less distortion (and guide or perception of the geospatial distribution of values). The Plate Carre map projection (the hideous default projection used by too many software packages) treats geographic coordinates (long, lat, radius) as if they were cartesian coordinates (x,y). The result, in this case, is that the US gets compressed along the N-S direction. Montana is large enough and doesn’t need to be exaggerated as it is in this display :)
If you’re using R and the ‘maps’ and ‘maptools’ packages, then take a look also at the ‘mapproj’ package.
Most choropleth maps are supported by ‘equal-area’ projections. Given the shape of the US (large east-west extent relative to its north-south extent), a conic projection method is appropriate. Nathan, if your data scientists are gonna make effective maps, then they need some background in cartography/GIS.
How revealing to choose green and greener to represent abortion and more abortion.
DaveW, what color would you have chosen?
(Given blue and red are usable.)
Sorry… Blue and read are NOT usable.
R code, please!
I agree on the projection problem.
One idea is that people travel to the nearby more “accessible” states for their abortion, especially if they live in less urban states. Examples include:
Wyoming -> Kansas/Colorado
South Carolina -> Georgia/NC
Mississippi -> Alabama/Tennessee, and
Utah -> Nevada.
Obviously, you would “go visit a friend” or “go on a weekend trip” to avoid being detected in smaller towns/rural areas where someone always recognizes you, likely while visiting the clinic.
This is a fascinating view, it would be great to so this data over time, perhaps decades? Great stuff!
I agree that it would be more impactful with a shorter time frame and overlaid with information about access to abortion services, for instance the number of clinics in a state.
Lotiat, Yellow? Gray?
Pingback: Science Etcetera, Moonday 20090615 | ideonexus.com
Pingback: digg » Blog Archive » Roundup: Third-Trimester Abortions May Continue in Kansas
Will we ever devise a standard way to ensure that we include ALL 50 states? Or is Roe v. Wade inapplicable in Alaska and Hawaii?
The question of abortions by residents of a state as opposed to abortions performed in a state is an important one. Many people travel long distances to get abortions in NY since it’s one of the few places in the US where it’s possible to get one if you’re late-term.
Also important is the availability of abortion-providers, which is some states may be in the single digits. You can’t get an abortion if you can’t GET an abortion.
Finally, the using the area of a state to represent it is, of course, misleading in terms of population, and therefore frequency of any human activity. Fight the non-normative norm!
Thanks for the discussion.
I’m looking forward to the tutorial — I want to start mapping in R myself!
“Many people travel long distances to get abortions in NY since itâ€™s one of the few places in the US where itâ€™s possible to get one if youâ€™re late-term.”
Actually, NY isn’t one of those places. Late-term abortions are heavily restricted in NY.
The key to understanding this map is that the data is from 1970-2005. Roe V. Wade was decided in 1973. Prior to Roe V. Wade, only two of the lower 48 states allowed abortion for non-health-threatening reasons — NY and WA. So during this 3-year head start, all of the (legal) abortions in the mainland US were concentrated in NY and WA. It’s likely women traveled from all over.