Skip navigation
All People > cfrye-esristaff > One Map at a Time

We recently shared a series of county-level maps showing the state of COVID-19 in the U.S. The maps show trends, days since last case, recent outbreaks, mortality, and state-level summaries. The maps are updated daily. 


Each day's build of the Johns Hopkins University data gives us the ability to compare the trends to any of the previous days. Here are some facts we learned today regarding the past week :


Today there are:

99.50% of Americans live in a county or territory that has had at least one confirmed case of COVID-19.

91.97% of Americans live in a county or territory with a spreading or epidemic trend.


  • 96 counties, 812,697 people (0.24%) have Emergent Trend
  • 1074 counties, 150,334,503 people (44.74%) have Spreading Trend
  • 690 counties, 158,682,332 people (47.23%) have Epidemic Trend
  • 649 counties, 18,556,404 people (5.52%) have Controlled Trend
  • 360 counties, 5,922,691 people (1.76%) have End Stage Trend
  • 274 counties, 1,682,521 people (0.5%) yet to confirm a case.


We are slightly better off today than we were a week ago:

  • 264 counties with 13,460,430 (4.01%) Americans improved their trend over the past week.
  • 383 counties 12,585,574 (3.75%) Americans worsened their trend over the past week.
  • Details for improved counties:
    •  35 counties with 327,281 people (0.10%) improved from Emergent to Controlled Trend
    •  64 counties with 4,706,559 people (1.40%) improved from Spreading to Controlled Trend
    •  14 counties with 4,431,529 people (1.32%) improved from Epidemic to Spreading Trend
      • Includes Santa Clara County, CA - 1.9m people 
    • 151 counties with 3,995,061 people (1.19%) improved from Controlled to End Stage Trend
  • Details for worsening counties:
    •  5 counties with 71,121 people (0.02%) worsened from Emergent to Spreading Trend
    •  5 counties with 44,136 people (0.01%) worsened from Emergent to Epidemic Trend
    • 107 counties with 5,146,374 people (1.53%) worsened from Spreading to Epidemic Trend
    • 176 counties with 6,149,324 people (1.83%) worsened from Controlled to Spreading Trend
    •  32 counties with 376,192 people (%) worsened from Controlled to Epidemic Trend
    •  52 counties with 732,089 people (%) worsened from End Stage to Controlled Trend
    •  4 counties with 49,041 people (%) worsened from End Stage to Spreading Trend
    •  2 counties with 17,297 people (%) worsened from End Stage to Epidemic Trend

Map Projections:

Posted by cfrye-esristaff Employee Jul 24, 2018

Today, a colleague asked me: "Do we have any base maps in ArcGIS Online that are close to the Hobo-Dyer projection, below? Or something similar that is equal area?"  Here was my response:


We have both Cylindrical Equal Area and Behrmann, so why would I use this one (the wiki does not give me a compelling reason)?


I make this point because there are better options for cartographic visualization (Mollweide and even Goode’s).  For Analysis, Cylindrical Equal Area is very good, except in extreme polar regions. However, the Add Geometry Attributes tool fully compensates for analysis by calculating geodesic area. I’ve done a comparison on a half degree fishnet and only found the only differences were in the two uppermost and lowermost latitude rows, and most of them were very close. If I were analyzing a polar area, I’d be using a Stereographic projection for visualization with that geodesic area option for analysis.


In general, I think a lot of the ‘ad nauseam’ projected coordinate systems were responses to lack of coordinate precision (pre-32 bit databases), and relatively slow computing technology. Today, my suspicion is we could get away with as few as twenty projected coordinate systems (the rest would generally fall under artistic license).


I would add that for analysis, equidistant projected coordinate systems are also vital for spatial analysis; and Equidistant Cylindrical is my preferred option.


For context, I have spent a good deal of my time in recent years working on global datasets, needing to deal with relative large areas and great distances, rather than regional or small study area geographic extents, where there are other proven options (national grids or state plane).





The Content Team at Esri has recently released a 2015 edition of the World Polulation Estimated Layer.  Learn more on the ESRI Insider blog, inc links to the new layers in ArcGIS Online. Check out the links at the end of the blog to get the technical details from the ArcGIS online content items.

The Content Team at Esri has published just over 200 image services into the Living Atlas in just over two years. Most of these services can be used as layers in ArcGIS Desktop geoprocessing tools and models.  Learn how here:


Use Living Atlas Image Services in Your Desktop Analysis | ArcGIS Blog

Several weeks ago we released some new global layers in the Living Atlas. Here is a link to the ArcGIS Blog describing two of these layers:


New Global Layers for Local Relief and Landforms added to the Living Atlas | ArcGIS Blog

President Obama has proclaimed September National Wilderness Month, in correspondence with the 50th Anniversary of the Wilderness Act. This morning an exhibit opened at Smithsonian’s National Museum of Natural History featuring photographs of the U.S. Wilderness System. The exhibit is in observance of this month's 50th anniversary of the Wilderness Act, and features touch-screen kiosks with two Esri story maps. Here are the online versions: Wilderness Forever: The Photographs and Wilderness Forever: Wilderness in Context. Also see #Wilderness50.


It is also pretty amazing to see how far Story Maps have come in such a short time. Could you imagine having to make a poster and use it to create this kind of awareness?


A little over a year ago a group was added to ArcGIS Online with some beta content. The group is called Landscape Layers, and it contained about 70 layers that followed a new model of content publishing. Since then the number of layers has more than doubled and the content is no longer beta--these have become fully supported esri-authored content layers.


You might wonder why these layers were new and continue to be special? There are two big reasons:


  1. The content in these layers supports landscape analysis, which underpins a big part of land use planning. Read the group's description to learn more.
  2. Unlike previous content, which only supported mapping, the landscape layers support both mapping and analysis. That means many of these layers could be used as inputs in geoprocessing tools and models in ArcGIS Desktop or online via REST and the various ArcGIS APIs.


The significance of additionally supporting analysis is, in my opinion, huge. It means not having to download data first. It means not fixing the data to make it GIS-ready after you've downloaded it. In many cases, but perhaps most obviously with the Soils layers, you do not have to learn how the data was published before you can make it GIS-ready, or use it. This makes ArcGIS and your work more efficient by leveraging the Cloud.


That said, the technology for delivering these layers continues to be a work in progress. Today, image and feature services can be the basis for layers that support both mapping and anlaysis, however both have limitations with respect to how much data can be analyzed.


Image services are limited by the number of images and pixels that can be used. Feature services are limited with respect to the number of features that can be queried, and when the result is more than 1,000 features, performance is quite slow. Thus, we felt we could be ambitious in publishing the image services such that 24,000 x 24,000 pixels could be analyzed. The descriptions for each of the landscape layers contains information about what that specifically means given the resolution of the data behind the layer.


For vector data, we chose to deliver dynamic map services, rather than feature services. This was for two reasons: 1) drawing performance, 2) avoiding the query performance issues mentioned above while the development team is working to address them. To support analysis of that vector data we created a geoprocessing service (Extract Landscape Source Data) that works in ArcMap, allowing you to specify a study area extent that you digitize or that comes from an existing layer, which is sent to the service and the result is a ZIP file containing a geodatabase with the extracted data. See the description for details about how much data can be downloaded from each of the supported landscape vector datasets.


Last, and definitely very important to us, is your feedback. Use the comments in ArcGIS Online for layer-specific feedback.

A week ago Esri released an update to our Terrain Image service on  A major aspect of this was adding a multi-directional hillshade function (this has it’s own content item ).  In ArcGIS Desktop, this service can be used with color ramps that you chose to have different tinting schemes.


But the important thing is that it achieves something I’ve wanted to see for a long time which was to effectively portray the character and definition of the terrain in the illuminated areas as well as the shaded areas.  In my opinion, this is one of the most exciting things we have done in quite a while.