Mapping North Korea

UPDATE – As of 1/30/2013 Google has added North Korea features to its maps.  Feel free to continue reading this post, because as of 5/18/2011, when this post was written, it was spot on. Just imagine that the embedded Google maps don’t have any features.  I really should have grabbed screen shots when I wrote this.

A week or so ago I read a story from Amnesty International that discussed the use of satellite imagery to measure the growth of  political prison camps in North Korea.  I thought it was great that these citizen geographers were using a resource such as satellite imagery to measure this troubling phenomena.  Then, the inquiring geographer in me wanted to know where these prison camps were within North Korea. A quick Google/Wikipedia search gave me some coordinates and I was able to find the locations on the map.  I then noticed something, that to me, that was way more interesting.

I zoomed out just a little bit and I noticed the total lack of cartographic markers on the map. I then zoomed out some more and there were still no place names, regional boundaries, city names, or physical features.  Test it for yourself, zoom in and out on the following Google Map.  Pretty amazing.

View Larger Map

Then I checked out the region around the capital city of Pyongyang, which is a city of over three million residents.  Still, nothing.

View Larger Map

The researcher in me then wanted to compare the cartographic representation of North Korea using other popular online map services.  I did a simple search for North Korea and Pyongyang using Bing Maps, Open Street Map, and MapQuest. The following results are very interesting.

Bing, owned by a gigantic corporation, which has used professional data providers for many years, has a much better North Korea map than Google, with cities, regional boundaries, and physical features labeled:

and Bing’s Pyongyang map had some minimal detail to it.

Open Street Map (OSM), an open source, user created worldwide map, was by far the most detailed.  Why? Because the OSM community created and published the data themselves.

View Larger Map

There is also far more detail in and around Pyongyang than Bing and Google.

View Larger Map

MapQuest (yes, they still exist) was a happy medium between Bing, and OSM.  There was detail in regards to the major cities, regional boundaries, and physical features.  There was also a clear difference in terms of  the cartographic details when one compared the cartographic features in the neighboring countries of China and South Korea.

Around Pyongyang the level of map detail was greater than Google and Bing, but less than OSM.  The user can make out details related to major transportation networks and some local features.

The overriding questions is, why would these three other web mapping providers map this information but Google wouldn’t?  I searched for a reason throughout the interwebs and I couldn’t find a legitimate answer.  I understand the North Korea is a closed society, run by a dictator that has a horrible human rights record. However, if this apart of Google’s “do no evil” approach to business then why are other countries with troubled histories, dictators, or in civil wars mapped?  Whatever the reason, there are a number of other sources for mapping data that are easily accessible.


Spring NEArc, Tuesday May 17th!

The annual spring NEArc meeting is tomorrow, May 17th at Smith College in Northampton, Massachusetts.  This is a great one day conference, bringing a variety of talks from education, public, and private sector GIS users.  I’ve spoken at a couple spring NEArc conferences and I have always enjoyed the people who make it out to western Massachusetts for the meeting.  The talks usually represent a good mix of GIS applications, techniques, and theory. So, if you are in the area, check it out.  Onsite registration is will be available on the day of the conference.

The schedule is now online. When you are there make sure to check out “Comparing Neighborhood Change in Connecticut 1934 to Present using Google Maps API” from my friends at MAGIC!


Upcoming GIS Conference – Harvard Center for Geographic Analysis

I will be attending Harvard’s Center for Geographic Analysis annual conference as an interested spectator this Friday and Saturday, May 6th and 7th.  This year’s conference is focused on the future of web-mapping and geo-collaboration.  Reading through the program it looks like there will be a number of very interesting talks and demonstrations.

From the program:

“This two-day conference will bring web-mapping experts from across the country to Harvard to share their knowledge and experience, and to envision what lies ahead. Speakers will introduce ideas, systems, tools, and visions, and present case studies and discuss challenges. Some will offer live demos and hands-on training. This conference provides a forum for geospatial technologists, developers, academics and end users to engage in dialog and help shape the future of geospatial technology on the Internet.”


I am interested in seeing what the world of online mapping is to people who aren’t like me, a trained GIS geek.  Because, let’s face it; many new and innovative web-mapping tools don’t come from geographers, but people who have a particular problem and a new way to solve it.  I am hoping to tap into the web-mapping mojo that will be in abundance at the conference and perhaps incorporate some of the ideas into my current projects.

It looks like registration is still open for the conference. If you are in the Boston area on Friday and Saturday and want to be on the cutting edge of web-mapping and geo-collaboration this could be a great conference for you!

Google Map V3 WMS Transparency Slider!

Gavin, from sent me a note the other day about some work he did in regards to developing a transparency slider for Google Maps V3:

“I’ve created a transparency slider control that works with Google Maps API V3, if there’s enough demand I’ll take some time and extract it from my code to publish as a nice neat example so others can use it too. See it in action at use the “contact” tab there to get in touch if you’re interested.”

The transparency slider works perfectly and the slider bar fits well into the layout of the page.  I hope to take Gavin’s work and create a stand alone example sometime over the next few days.  In the meantime check out if you have some free time!

View Larger Topographic Map

Everything Should be Mapped – Somerville Tax Liens

I came across an interesting dataset the other day while reading the Boston Globe (online of course, does anyone buy actual newspapers anymore?).  The story, originally from Ward 5 Online, was discussing the number of properties in my current hometown, Somerville, Massachusetts, that have tax liens over 10,000 dollars!  The article provided a table from the City that included the address, owner name, and amount in back taxes for each property.   The story reported that this data was made available by the City’s aldermen through a request regarding derelict buildings that were shedding bricks (yikes!).

After I read the story I started to search for some of the properties to see if they were in my neighborhood.  I quickly realized that this data would be better utilized if it were in a map.  Thanks to Google Fusion I was able to quickly modify the table and map the addresses and back tax balances.  In the map I created I only included the amount owed, the location, the status (many of the properties have been taken over by the city), and the number of years that the property is late on their taxes.

Back Taxes Legend:  10k to 15k 15k to 20k   20k to 25k   >25,000k

This data can be further analyzed with a number of other free datasets including census, income, or home sales data.  One could then perform any number of analyses to see if there are any spatial patterns in regards to these large tax liens throughout the City (aka, future blog idea).

Also, this map is not perfect.  We all know that there are limitations to mapping with Google, especially in reference to thematic mapping, but overall one cannot complain about the speed and efficiency of creating a simple, but effective map with Google Fusion.  One item I would like to see from Google would be an embeddable legend.  Esri’s JavaScript API can do it, why can’t Google’s?  Now, we can have a discussion about how Google is collecting all of this data, having us create it for them(for free), and then doing who knows what with it, but that is for another day (aka, another future blog post).

Over the next couple of days I will take my dataset over to GeoCommons and work with their tools to create a better map and perhaps do some analysis (aka, another future blog post).

Until next time…

Have fun at AAG!

The annual AAG meeting is being held this week in Seattle, Washington (April 12-16, 2011).  First of all, what is the AAG?  The AAG is the Association of American Geographers, an organization that represents both academic and professional (but mostly academic) geographers that has several thousand members from both the United States and abroad.

The annual AAG meeting is biggest geography meeting of the year, and I should know.  I have been an AAG member for close to ten years now, and I have presented at the AAG meeting several times.   The AAG meeting is like the discipline it represents; very diverse and multidisciplinary.  At the meeting you will have talks about feminist perspectives of geography, migration studies, land use/land cover analysis, environmental studies, qualitative analysis (I don’t go to those talks, I’m a numbers guy), location analysis, cartography, GIS and its thousand iterations, geomorphology, regional science,  radical geography (yes, that is a real sub-field of geography), geographic education, and much, much more.

Unnecessary side note: At the meeting in San Fransisco my friends and I saw a bait car get pulled over outside of our hotel.  It was like watching Cops, but from the fourth floor of a hotel.

Anyway, the AAG meeting is so big that people often get lost in the meeting schedule and end up missing out on some really good presentations.   At the meeting there are several thousand attendees, and thousands of presentations.  Why are there so many presentations?  Well, the AAG meeting is non-refereed, meaning that you submit a talk and in the inner workings of AAG someone will assign you to a time slot, hopefully in a session with somewhat related talks.  This process is aided by the many special interest groups that help organize the vast number of talks into a working schedule.

In my experience it is best to connect with other geographers (or interdisciplinary folks who secretly want to be geographers) who are in the same research field as you to form a session.  This method ensures that the people in the room during your talk know what you are discussing and hopefully find your work interesting.

Second unnecessary side note: At the Las Vegas meeting my friends and I went to Red Rock Canyon to watch the sunrise after a night a enjoying the town.  Definitely a place to visit when in Las Vegas.

It does takes some skill to navigate the schedule.  I believe you have to search the schedule for specific topics.  Using a general keyword like ‘GIS‘ will return a million different talks. Also, when searching for talks I would stick to sessions that are sponsored by the special interest groups you are a member of.  My biggest pet peeve about AAG (besides never having a place to sit in the lobby of the conference center) is to end up in a talk that has an awesome title and a good abstract but has painfully horrible content.

Alas, I will not be going to this year’s meeting.  Like many other geographers I will be attending a number of smaller and more technically specific meetings and one day conferences instead of the annual AAG meeting.  I tend to get more out of a smaller conference or one that is very technically oriented, rather than a large meeting where you are elbow-to-elbow with a 1000 undergrads in the book fair.

If you are in Seattle for the AAG meeting have fun.  While you are there check out the sites, catch up with some old friends and get a free drink at another school’s alumni party, get a good meal or two, and go watch a Mariners game.

Keyboard Shortcuts in ArcGIS 10

I’m always looking for ways to make my workflow more efficient, from creating scripts to automate a process or trying to simplify models so that they are both effective and efficient.

What about making the day-to-day use of ArcGIS faster?  Many of us create our own toolbars or custom buttons to make workflow faster, but what about using keyboard shortcuts?  If you do a lot of programming, or spend a lot time using any piece of software you want to know the available keyboard shortcuts.

We use keyboard shortcuts all the time from ctrl+c, ctrl+v, ctrl+s and many, many others.  What about keyboard shortcuts in our favorite program, ArcGIS?  Yes, they exist, and many of them are pretty helpful.  The basic ones for copy, paste, and save all exist, and there are many more including shortcuts for toggling all the layers in the table of contents, working in a table, or editing features.

Where can one learn all about these shortcuts?  Well, Esri has published a document on this exact topic that I find very useful.  There are a number of keyboard shortcuts and a number of other general tips that will benefit the novice or expert user.

Arc Shortcuts

What about tips and shortcuts for other tools that GIS users access all the time?  Well, there are number of tips for the field calculator, however, this tipsheet is a little old, but still relevant.

Know of any shortcut lists?  Post a link the comments section.  I’m sure someone will find it helpful.


Massachusetts Census and Google Fusion

2010 Massachusetts Census Results by Town  – Google Fusion

Population | -5,000 | 5,001 -30,000| 30,001-75,000 | 75,001-125,000|+125,001

About a week ago I read a post from James Fee about a really impressive mash-up displaying the 2010 US Census redistricting data.  The mash-up came from WNYC displaying NYC’s 2010 US Census redistricting data by census tract. The development of the mash-up was explained in detail in a blog by John Keefe.

The map authors used Google Fusion to serve their data.  As we know Fusion allows for data with a spatial component to be easily mapped in Google Maps, but polygon mapping is limited to large geographic extents, like states or countries.  This limits users who may want to map data that doesn’t fit this mold.

The authors of the WNYC map needed more geographic detail, so they imported a number of census tract shapefiles into Google Fusion using a tool called Shape to Fusion.  They then merged several tables with the spatial data and then visualized it using Fusion’s visualization tools.  Shape to Fusion provided a great work around.  This online tool loads a zipped shapefile into a table in the user’s Google Fusion account, creating a geometry column.  The geometry column can then be mapped in Google Maps.  Brilliant!  The concept of storing geometry in a database isn’t new, but Shape to Fusion’s capabilities in the Google Fusion interface are slick.

So, after reading John Keefe’s blog I wanted to try it our for myself.  First, I went to MassGIS and downloaded town boundary data that included population counts from the 1980 to 2000 census.  Next, I downloaded the 2010 town census data from the Massachusetts Secretary of State.  After I loaded the 2010 census data into my Google Fusion account I used the Shape to Fusion tool to load the zipped shapefile into my Google Fusion account.   I merged the spatial data and the census data based on the town field and presto, web maps.

Following the advice of John Keefe‘s blog I created a separate table for each layer I wanted to map and the results are below.  Now, I don’t have the bells and whistles that the WNYC map has yet, but I was able to generate four embeddable maps in minutes that fit well in a blog format.  In fact, it probably took me longer to write this blog than it did to create the maps.  When I get some free time or when SF1 comes out I’ll build a more robust mash-up using this technique.

Here are the maps.  The user can click on a town a get some basic information as well.

1980 Massachusetts Population

Population | -5,000 | 5,001 -30,000| 30,001-75,000 | 75,001-125,000|+125,001

1990 Massachusetts Population

Population | -5,000 | 5,001 -30,000| 30,001-75,000 | 75,001-125,000|+125,001

2000 Massachusetts Population

Population | -5,000 | 5,001 -30,000| 30,001-75,000 | 75,001-125,000|+125,001

2010 Massachusetts Population

Population | -5,000 | 5,001 -30,000| 30,001-75,000 | 75,001-125,000|+125,001

Get mapping!


I love models

GIS models, that is…


I get to do fun things with GIS, like develop models that handle and process millions of records.  This model is used to clean a number of spatial variables for input into another analytical model.  This awesome monstrosity contains:

  • 205 million input xy points read from SQL tables that are processed into over 500,000 polygons
  • 500,000 polygons merged and manipulated into about 200,000 polygons
  • Over 20 different spatial processes (merge, clip, field calculations, projections, spatial stats, etc…)
  • Database driven model, using a mix of SQL tables, SDE datasets, and File Geodatabases
  • Four custom python scripts, some of which are repeated several times
  • The output? Two text files, which are generated from python scripts.
  • Output used in another set of spatial models

The model took about a week to build and calibrate.  The model itself runs in about 30 hours on a four core, 64 bit machine and produces close 20GBs of data from about 4GBs of input data.

Anyone can build a mapping website (including me!), but a true GIScience geek lives on this stuff.  I love spatial analysis.

2010 Massachusetts Census Results – Dual Maps Mash-Up

Taking a page from my friends at MAGIC and the Connecticut State Data Center, I’ve put together a simple dual-map Google Map mash-up of the recently released Massachusetts 2010 census data using the Google Map V3 API (see their example here).  The release of the 2010 Massachusetts census did have its problems, but now the correct data are available for download.  The census data are displayed using a dual-map approach. The top map displays the 2010 town-by-town population count while the bottom map displays the percent change between the 2000 and 2010 census.

Click here to visit the interactive maps.

Dual Census Map

Being a dual-map the zoom and pan controls are mimicked on each map.  For example, if the user zooms or pans on the top map, the bottom map reacts accordingly.  The user can also click on a polygon and a small info window will open that displays some basic census data.  Over the past ten years population in Massachusetts has grown slowly, with increases in the towns of central Massachusetts.  However, population has declined in many western Massachusetts towns.

The code for this page can be easily replicated and used for any number of other states, counties or towns. The user would have to create their own kmz or kml file and modify the javascript and html for their own needs.  Unfortunately, Massachusetts is a tough fit for a webpage due to its shape.  In this example much of the screen is dedicated to either ocean or other states.

*Update – I changed some of the code to position the maps and the legends so that they look better when the user shrinks the screen.  The KMLs may also load slowly, depending on your connection speed.

Happy mapping!