I am having trouble with the albersusa library. I am getting errors saying it is not available in R version 3.3 or 3.4. I could use some advice. Thanks!

Hello Sean,
The albersusa package is available on Github, make sure you have devtools installed and run:

devtools::install_github("hrbrmstr/albersusa")

Thanks! I figured that part out. My current issue is trying to map my data. So, I have a list of states, counties, and revenue. I want to color code based on the revenue. When I track my data though, I get the line data.ds<-left_join(cty_sf,ds.counties,by=c("state","county")) . At this point, my revenue data is gone and i cannot figure out how to get it back. Suggestions?

Sean,

If there is no data at all for any of the counties after the join then I would maybe suspect that your county/state data types are not the same. Look into those, to see if your cty_sf or ds.counties data types are both characters, or both factors.

That is where I would start!

That is my script. My original data has a column for County, State, and Revenue. I cannot figure out what is wrong.

So when you run class(ds.counties$county) & class(cty_sf$county) they are the same? (The same goes for state)

Actually they are different. Not sure why or how to fix it. I'm a bit out of my comfort zone at this point lol : )

No problem, we will hopefully get you there. Can you tell me which one is which? cty_sf$county should be 'character' so if you run:

ds.counties$county<-as.character(ds.counties$county)
ds.counties$state<-as.character(ds.counties$state)

Then try your join. For future reference I imagine that the ds.counties may be 'factor' data. If you read in a csv or excel you can set a parameter within that call, 'stringsAsFactors = F', which would make those strings as characters.

Yes, cty_sf$county is a character column . I ran those two commands, but it hasn't seemed to change my result. I already had the stringsAsFactors set to false when i imported my csv. Perhaps I am not understanding how all this works. There are 998 rows in my csv. However, when I look at the data.ds (after the join) There are 3142 rows.

Ah, I see, so the left_join command is joining the data you have, data.ds to the cty_sf. It will retain all the data from the cty_sf and all the matches it has from your data. IF you would like the join to join on only matches you can switch the join around, thus:

data.ds<-left_join(ds.counties,cty_sf,by=c("state","county"))

Gets you only the data that matches your ds.counties data. One caveat in doing this though is that when you go to make the map there will be lots of counties that will not have any 'fill' or color because the data is not there.

That is good to know. I did not realize that. That has my data frame correct. However, it is not plotting anymore. I had an [incorrect] plot before, but now there is nothing plotting.

Ah. Well I am getting to the extent of being able to help without seeing the script, your instance, errors and the data. I would suggest inspecting the data after the join, some column names may have changed since you switched around the join. If you run, nrow(data.ds) and it comes up as 0 that means there is no data so the join did not work. So you can fiddle with that. If you would like more assistance visit my consulting page, and put your contact info in the contact form, and we can discuss it further.

could i have a link to your consulting page?

http://www.austinwehrwein.c...

Thanks for the tutorial. When I use your code, it cant find the theme_awhstin. Do I need to update to a newer version to get this?

Hi Brad, thanks for reaching out. Unfortunately theme_austin is my own custom theme. But if you download the hrbrthemes package from CRAN there is a much better option called theme_ipsum!

Thanks for the tip - that looks great!

hi
1.how to cut specific map ?
2. how to add name inside of polygon?
3. give me suggestion what things needs to do for good mapping?

Awhstin, thanks this has been so helpful to me. Two quick questions, any way to add the state lines to the second map? I tried adding geom_polygon(data=us,aes(x=long, y=lat, group=group), color='dark gray',fill=NA, alpha=.35) to the last block of code but that didn't work. 2nd, any way to manipulate the colors?

Thanks again so much for your help.

Hi Tim, thanks for reaching out and glad you like it. I think you can add lines fairly directly in the geom_sf() section by adding color and you can adjust the alpha and line size there. The colors are both handled by the scale_fill_viridis and scale_color_viridis which are color blind friendly palettes you can read more about via this link https://cran.r-project.org/...

Please let me know if you have other questions.
Austin

Thanks for the response. When I try adding geom_sf(color = 'gray') it gives me lines around the county. Do you know how I could make the outlines of the states instead like you have in the original map?

Hi, I am getting this error with GEOM_SF().

Error in geom_sf() : could not find function "geom_sf"

I have loaded all related library, not sure why I am getting error.

Afternoon Nikhil,

I believe it is because you are not using the dev or github version of ggplot2 (which I mentioned in that blurb just before trying the final plot) which has geom_sf(). Try running (but make sure you have devtools installed):

devtools::install_github("tidyverse/ggplot2")

Thank you for this guide! How do I add AK and HI to this code?

Hi, if you look at the original data exported, it includes AK, HI and even some others. The reason they do not show up is because we merge via this line 'choropleth <- merge(county_df, fm, by = c("state", "county"))' the county map with the farmer's market data. This county_df is built from the 'map_data('county')' line. This data only has the lower 48. I would suggest visiting https://rud.is/b/2016/03/29... to see how to implement the Albers projection. I will, when I get the chance, work on an addendum to this post that includes the Albers projection.

https://uploads.disquscdn.c...

Austin,

Thank you for your reply. Yes, neither state data nor county data has AK/HI. Results from both of the commands below return 49 states. I will explore Albers projection next. Thank you for the tip.

us<-map_data('state')
unique(us$region)

county_df <- map_data("county")
unique(county_df$region)

Min

Min,

Here is a little bit of code I think you can add to what you currently have to get all 50 states. (Be sure to have the albersusa package downloaded off of github).

us<-data.frame(counties_composite())
names(us)[7:8]<-c('statename','state')
names(us)[4]<-c('county')

#put the pieces together
choropleth <- merge(us, fm, by = c("state", "county"))
cmap <- fortify(counties_composite(), region="fips")

gg <- ggplot()
gg <- gg + geom_map(data=cmap, map=cmap,
aes(x=long, y=lat, map_id=id),
color="#2b2b2b", size=0.05, fill=NA)
gg <- gg + geom_map(data=choropleth, map=cmap,
aes(fill=count, map_id=fips),
color="#2b2b2b",size=.05)+scale_fill_viridis(option="magma",direction=-1)+theme_map()
gg

Which should get you this, which is straight from the example on the vignette. Hope that helps,
-Austin

https://uploads.disquscdn.c...

Wow, thank you Austin. It works!

Min

Any other updates to this?
Once correction: changed first ggplot command to:
ggplot(fm,aes(longitude,latitude)) +
geom_polygon(data=us,aes(x=long,y=lat,group=group),color='gray',fill=NA,alpha=.35)+
geom_point(aes(color = count),size=.15,alpha=.25) +
theme_bw()+
xlim(-125,-65)+ylim(20,50)

Also, getting error at: > fm$county<-tolower(fm$county)

So for the county merge I left out this section, which is now fixed on the website:

fm.zip<-aggregate(data.frame(count=fm$FMID),list(zip=fm$zip,county=fm$County),length)

Thanks! I wonder though: if I want to map to a zip code polygon, what's best? I have had a lot of trouble using the choroplethr package. Do you have an example that maps zip code level data? Thanks!

Sure no problem, sorry about the error in the first place. As far as the choroplethr package, this tutorial does not use that, I just named an object in R 'choropleth' which can be confusing. The result should be like the last image, where it maps the counties as polygons, colored by the value. Let me know if there is something else you mean.

The last/bottom map on this page is a USA map that appears to show counties. Can a method be used (not the choropleth package) to show a map of ZIP codes? If so, how?

Oh, terribly sorry I misunderstood your question, I have not looked directly into the choroplethr package nor have I tried to map by ZIP codes. I won't be of much help to you there. I have read though that trying to do ZIP codes on a US scale may not render nicely or quickly. Most people suggest mapping by the ZIP if you are looking at a subset of data. If that is what you want to do, I think finding a shapefile of ZIPs for whatever area you are looking for wouldn't be too difficult! I hope this helps :)

Hello, terribly sorry for the errors, I will take a look and post back

Please ignore my previous comment about ggmap and R 3.3.0, thanks.

It helps to use the correct spelling :)

Hi James,
in your response you're referring to R 3.3.0. It seems that the package ggmap does not work with 3.3.0 (yet?). Do you have a work-around (i.e. how to use get_map with R 3.3.0)?

Hi,
Im trying to work with this (new to R) and finding that the zip codes package is not available for R 3.1.2. Is there a workaround for this?

Hello Zoe,

Thanks for your question. I don't believe there is an option for the zipcode package for that old of a R version. The most recent version of R is 3.3.0, here is a link to update your R if you can, if that is not an option we can try to work through something, here is the windows download link
https://cran.r-project.org/...

Reblogged this on Austin+Wehrwein and commented:

Here is a re-blog of one of the most popular tutorials I have done, with R and a creating a map from a csv of zipcodes.

Is there any way you can demonstrate drawing a map of a specific state?
I've tried using subset method for a specific state, but the size of the map stays the same and it doesn't really zoom into specific states that I chose.

Thanks!

Hello Martin, thanks for reaching out. The subset is a good start and should work if you subset both the 'state_df' and 'county_df' what I would suggest is simply adding your state to the criteria for creating the data for the maps like so;

county_df <- map_data("county","nebraska") #or the state you want
state_df <- map_data("state","nebraska")

Awhistin, thank you very much for the reply.

I am not too sure how I should go on about what you mean by subset both 'state_df' and 'county_df'.
my current code looks something like:

us<-map_data('state')
#specificss <- subset(us, region %in% c( "virginia", "maryland", "north carolina"))

ggplot(fm,aes(longitude,latitude)) +
geom_polygon(data=specificss, aes(x=long,y=lat,group=group), color='gray',fill=NA,alpha=0.35) +
geom_point(aes(color = count), size=0.15, alpha=0.25) +
theme_grey()+
xlim(-125,-65)+ylim(20,50)

The problem is, it still shows other states even though I have subsetted them out; for example, the map doesn't seem to zoom into those three states that I want.

Thank you very much in advance

Hi Martin, are you also subsetting the 'fm' data for your states? What I mean is you have also subset your 'fm' data for those states as well since the geom_polygon is only plotting the data to fill the map but the underlying base from ggplot(). So try something like this;

subset(fm,state %in% c('MD','VA','NC'))

Hello Awhstin,

Sorry, I was doing it all wrong. I was still working on the first part of your code (code that is for the very first picture) and trying to figure out how I would be able to incorporate subset states and counties lol. I figured out after looking at the second part of your code! thank you very much!

By the way, is there any way that I would be able to draw the map by purely using given Zip codes other than Longtitude + Latitude? I know zipcode package will be able to pull long + lat info just by looking at the zip code.

Best,
Martin