Skip to content

John Snow's cholera data/map: a more sp-ish way?

4 messages · Michael Friendly, Roger Bivand, Stéphane Dray

Original
Michael Friendly
# 
Hi

In a new package, HistData, on R-Forge, I have a collection of 
historical data sets, some map-related,
including the map and data related to John Snow's famous map of the 1854 
cholera outbreak in London.
https://r-forge.r-project.org/projects/histdata/

I have two examples in example(Snow) that re-create the map of deaths, 
one showing Theissen polygon
boundaries around the water pumps, the other showing  bivariate density 
contours, but they are entirely
done using base graphics and ordinary data frames.  I'm wondering if 
anyone is willing to help me show
a more sp-ish way to organize the data and map info and make these and 
other plots.
More generally, I would welcome any other contributions to this project.

-Michael
Roger Bivand
#
On Fri, 30 Oct 2009, Michael Friendly wrote:

            

      
      
      
    

        
Michael,

This is good news. I'm sorry that I didn't respond to your message about 
the Guerry dataset from almost two weeks ago, I ought to have done. I'll 
try to get back to you on both that question:

https://stat.ethz.ch/pipermail/r-sig-geo/2009-October/006673.html

and this one. I encourage others to join in, because the R-forge package 
has the potential to become a valuable resource.

Best wishes,

Roger

            

      
      
      
    

        

  
    

      
      
    

  


  


      

    

    

      
      

        


  

        

      Stéphane Dray
    

    

      
      #
    

  


  

      
Hi Michael,

I tried to play with the streets and transform them into a spatialLines by:

l1=split(Snow.streets[,3:4],as.factor(Snow.streets[,1]))
Ll1=lapply(l1,Line)
Lsl1=Lines(Ll1,paste("Street",1:length(Ll1)))
Snow.streets.sp=SpatialLines(list(Lsl1))
plot(Snow.streets.sp)

For the pumps:

spp=SpatialPoints(Snow.pumps[,3:4])
Snow.pumps.sp=SpatialPointsDataFrame(spp,Snow.pumps[,1:2])
plot(Snow.pumps.sp,add=T,col='red',pch=20)

For the deaths:

Snow.death.sp = SpatialPoints(Snow.deaths[,2:3])
plot(Snow.death.sp,add= T, col ='blue')

For polygons, I think that it should be a little bit more complicated.. 
but I have no time today to have a look at it.

Is it correct Roger?

Cheers.

        
Roger Bivand wrote:

            

      
      
      
    

        

  
    

      
      
    

  


  


      

    

    

      
      

        


  

        

      Michael Friendly
    

    

      
      #
    

  


  

      
Thanks, Stephane

This has most of the pieces.  The Snow.polygons aren't really polygons 
-- just a set of line segments
defining the boundaries of the Theissen/Voroni polygons giving the 
boundaries of the regions of deaths
"nearest" to each pump.  I was just including here what I got from Waldo 
Tobler, but there may have
been something more sophisticated than just Euclidean distance used here.
There may be a more sp-ish way to calculate these directly, or do 
something similar, from the Snow.deaths
and Snow.pumps.  The visual goal is to show the concentration of deaths 
around the Broad Street pump.

For the HistData package, I'd like to simplify this a bit, complete it, 
at least with a title and mapscale,
and consider if I should include some Spatial objects in the package 
directly or else leave this to the
example.

Why don't I get a title using main= below?
# streets
l1=split(Snow.streets[,c("x","y")],as.factor(Snow.streets[,"street"]))
Ll1=lapply(l1,Line)
Lsl1=Lines(Ll1,paste("Street",1:length(Ll1)))
Snow.streets.sp=SpatialLines(list(Lsl1))
plot(Snow.streets.sp, main="Snow's Cholera Map of London")

-Michael

        
St?phane Dray wrote: