Digging Up Missing Cemeteries

by LB2019gis

Part 1: The Texas Missing Cemetery Project 

"Personally, I don't really like cemeteries, they give me the creeps but somehow, I got hooked on editing them for The National Map Corps (TNMCorps)! This is the story of how I ended up editing about 14,000 cemeteries over the last 3 years. When I started as an editor in early 2019 my first project was to tackle all the remaining unedited (red) points in Nebraska. I took this on for several reasons. First there weren't that many points needing editing. Second, I have some personal familiarity with Nebraska. And third, I was just getting started so exposure to all the different structures was useful.  

"After Nebraska I set off to do the same in Kansas but there were a lot more unedited points, so I limited my focus to unedited cemeteries and from there took on unedited cemeteries in Oklahoma. When I hit Texas, I immediately noticed that there were just a lot fewer cemeteries than there had been in the states to the north. In fact, the cemetery density seemed to be about 1/10 that of the previous states. I couldn't think of any particular reason why Texas would have far fewer cemeteries than Oklahoma. Maybe since "everything is bigger in Texas" they just had fewer but much, much larger cemeteries. Or maybe Texans got buried elsewhere. Or maybe they preferred columbaria or other non-cemetery treatments. None of these seemed like good explanations especially when I took a look at the Historical Topographic layer and saw how many cemeteries were labeled or marked on there but were not included as points in the TNMCorps editor. So, the explanation seemed to be that the information source used to populate cemeteries in the TNMCorps editor was just less complete for Texas than for the other states. I wondered if this was a situation that could be fixed.  

"I did a bit of on-line searching and discovered that the Texas State Historical Commission (THC) has sponsored a cemetery recording project for several years and their geographic database was available for download. As of January 2020, it contained roughly 12,200 cemeteries compared to about 2,700 marked in the TNMCorps editor. While a great source for the missing cemeteries, in consulting with the TNMCorps staff, we decided it didn't make sense to directly import these points into the editor since that would create duplicates of those already in the database. I needed to figure out an efficient method to get those "missing" 9,500 points into the TNMCorps editor and "Missing Cemeteries 1.0" was launched. And I had guaranteed volunteer job security for the next couple of years!  

"Fortunately, I have a background in GIS (geographic information systems) and access to software, so I created a printed map of each county in the Texas Historical Commission (THC) database with labeled cemeteries, major roads, and large towns. These maps became my guide for adding cemeteries to The National Map. Since the THC database didn't include much additional information, I turned to Find A Grave (a web site TNMCorps also considers an authoritative source for cemeteries). My printed county map informed me about where a cemetery should be, a look at the TNMCorps editor told me whether it already existed in the database, and Find A Grave provided an authoritative source to be able to add the cemetery if it didn't exist. It took 18 months to work through all 254 Texas counties after which I'd edited some 8,900 points between new additions and points requiring editing (including approved points with no edit history) that were already in the structures database. I'm now working my way through other states tracking down "missing" cemeteries." 

Sources/Usage: Public Domain. View Media Details

Part 2: The Missouri Cemetery Project Background

During the Texas work, there were TNMCorps cemetery mapping challenges in southeastern Missouri and western Tennessee. So, after Texas, I moved on to Missouri where I edited 4,300 points of which 1,700 were additions to the database. When I finished up Texas (thinking it was a special case of having so many missing cemeteries), I looked at these two states as well as others to see if any popped out as needing significant cemetery editing, not necessarily just "missing cemeteries". I decided to work on Missouri since the challenge had left a good part of the state with a high density of unedited cemeteries and cemeteries needing peer review. I planned to do what I could to gather up any "missing" ones as I went. Missouri didn't have a cemetery recording project like Texas, so I needed a different approach to detect the missing cemeteries. "Missing Cemeteries 2.0" was launched and that is the method I'll describe here. 

Part 3: The Missouri Cemetery Project Method

As I did in Texas, I made a printed map of each Missouri county. This time instead of using an outside data source I used the TNMCorps data. I set the TNMCorps editor to show only cemeteries and then zoomed out (usually level 11 or 12, depending on how large the county was) so the cemeteries appeared as colored circles rather than cemetery icons. The "National Map Base Layer" provided major roads, towns, and water without being too cluttered. I then did a screen dump of the TNMCorps display and pasted the result into a PowerPoint slide. I cropped and sized the pasted image to fit on a 8.5x11 sheet and printed the map.  

Before beginning the next steps, I found it was important to clear my internet cache. While working in a county, browsing history needed to be retained throughout but cookies and cached images/files were cleared as needed.  

I used Find A Grave's county-level cemetery list as my "source" data and my goal was to add as many missing cemeteries from the Find A Grave list as possible and to be as efficient as possible while doing it. This was my approach: Zooming in to the TNMCorps editor so the cemetery symbols flipped back to icons from circles (zoom level 13), I selected each approved (yellow), advanced edited (pink), and peer reviewed (blue) icon and found that cemetery in the Find A Grave list. In Find A Grave, I opened the cemetery entry and then closed it with no action. This changed the font color of the cemetery name in the Find A Grave list from blue to brown. I marked off each cemetery on my printed map to keep track of those that were "done" and those that were not. I then selected each unedited (red) and peer review needed (green) point and edited or reviewed them as needed using Find A Grave as one authoritative source (see past TNMCorps newsletters and guides for authoritative source requirements). I used the Historic Topographic layer, Google Street View™, and satellite imagery as supplementary sources.  

To save some work later, when I was looking at the Historical Topo layer for the focal cemetery (zoom level 15), I'd look for any additional nearby cemeteries on the topo that could be added and would add them after I'd edited the red or green point I was working on. At this point, all the cemeteries on my paper map (and thus all the cemeteries in the TNMCorps editor) had been accounted for. The list of counties in Find A Grave had brown (those that were on my paper map or added) and blue (those that weren't on the map and thus not in the editor yet) entries. The good news here was that often the majority of the cemeteries in the Find A Grave list had been flipped to brown, so the list was now a lot shorter to work through. Note that if the browser cache was cleared of Browsing History during this process, the brown entries would turn back to blue. Also, by having looked at TNMCorps cemeteries across the county, I had become familiar with the location of the major roads, towns, and water features, giving me a sense of geography that became useful in the next part of the process. 

Now, the task got harder, or at least slower. I started at the top of the Find A Grave list and visually filtered the remaining blue entries for number of burials (e.g., >25-ish), and likelihood of being able to "see" the cemetery in Google or ESRI imagery (e.g., did the cemetery photo, if any, show more than just weeds or dense trees?). I then selected each list entry in turn and attempted to add it to the editor. I used the Find A Grave map display and coordinates to find the location in the TNMCorps editor where the cemetery would be located and panned Google Maps to that same location. If I could confirm that the cemetery was there (labeled on the Historical Topo, visible in Google imagery, etc.) I added it to the map with a comment supporting the entry. Since the Find A Grave list was alphabetical, there could be a lot of jumping around in the county which was where having "learned" the county geography earlier was helpful. My cutoff of looking only at Find A Grave entries that had more than 25 burials was a bit arbitrary, but I found that smaller cemeteries were just too hard to detect. Similarly, cemeteries that lacked a photograph or which had non-descript photo content were virtually impossible to do much with. 

Part 4: The TNMCorps Cemetery Analysis  

As I worked through Texas and Missouri, I noticed that I was able to populate the TNMCorps editor with only 50-60% of the cemeteries in the Find A Grave county lists. No surprise since I skipped the small and poorly recorded cemeteries in the lists. However, this tentative "rule of thumb" gave me a way to find out if any other states had "missing" cemeteries. The TNMCorps staff was able to give me a list of all the cemeteries currently in the database (a total of 166,563 in the 50 states plus the District of Columbia and Puerto Rico) so I compared state-level counts to the number listed in Find A Grave. 

Assuming that Find A Grave provides a consistent and reasonably comprehensive listing of cemeteries across states (which may not be true given the volunteer nature of the information posted on the website), then TNMCorps has excellent coverage of cemeteries in Iowa and Maine where 94 and 95%, respectively, of the cemeteries in Find A Grave are present in the TNMCorps editor. At the opposite end of the spectrum, North Carolina is "missing" around 18,000 cemeteries (TNMCorps has 3,600 and Find A Grave lists 36,100). In answer to my original query as to whether Texas was atypical in the number of "missing" cemeteries, in addition to North Carolina, both Kentucky (17,800) and Virginia (12,400) had a larger number. In all, there may be about 91,000 missing cemeteries that could potentially be added.  

If any editors are interested in working on cemeteries but don't want to take on the "missing cemetery" task, there's plenty of other low hanging fruit! The Western Tennessee cemetery mapping challenge addressed the west side of the state but there are still over 4,700 unedited (red) and 1,700 need peer review (green) points. Mississippi has 2,600 red and 1,100 green cemeteries. While Ohio has a high percentage of cemeteries from Find A Grave, it also has 2,900 approved (yellow) cemeteries without edit histories. South Carolina has almost 1,800. West Virginia has almost 1,200 cemeteries that need peer review (green) that also have no edit history. Georgia (3,000), New York (3,300), and Indiana (2,700) have large numbers of cemeteries needing peer review (green). 

Sources/Usage: Public Domain. View Media Details

-----

It’s likely that some of the numbers reported in this article have changed since it was written, however it still serves as solid guidance on where to search for missing cemeteries, if you’re so inclined to seek them out! 

The TNMCorps staff would like to sincerely thank LB2019gis for the tremendous and extremely thorough work they’ve done. Our volunteers are what makes TNMCorps so special, and we are so lucky to have dedicated folks with a passion for geography and maps put so much time and effort into making USGS map products and services better! We truly couldn’t do it without you. You rock!