Jump to content


Photo

Reverse Geocoding - to County or Not to County

mapping geocoding place list

  • Please log in to reply
3 replies to this topic

#1 Vyger

Vyger

    Advanced Member

  • Members
  • PipPipPip
  • 3144 posts

Posted 28 February 2018 - 07:29 AM

I can't claim original thinking for this post as it essentially came from another thread below where Jerry Bryan suggested Countries having a number code rather than a text description. One problem I would see with number coding would be that there would always be variations and lack of agreement although it would be beneficial in the international market. The numerous variations in text entry of Places we live with today is beyond solution or any agreement in my opinion so maybe we are looking at this all wrong.

 

http://forums.rootsm...to-place-names/

 

So the geocoding (pin in the map) is the best definition for me of where an event took place, no worries about modern or historical, no consideration of administrative boundaries, boroughs, electoral divisions or the like. If your pin in the map is right beside mine in the same time frame then there could well be an association! It doesn't matter what you call the Place, whether you include or exclude the word "County" include or exclude the Country, record the Country and "United States of America" or "U.S.A" the pin is in the same Place.

 

If you lift geocoding from an online map and then search for it in Bing or Google you will be returned with the location in text, Reverse Geocoding. Of course the returned text will not suit how many users like to see the text string but the geocoding is the accurate and mathematical representation of where the event took place and the most useful for proximity calculations of possible associated events.

 

"In a perfect world" this Pin in the Map and the event date would be translated by "County Check" to return a historical Place Name text where appropriate and if desired.

 

If a user is a national of Spain, China, Germany or other then the Reverse Geocoding process will return the place text string in their own language if the correct reference source is used.

 

So maybe we have been doing it all wrong, maybe in a lot of cases we should be sticking a pin in the map and not endlessly deliberating over what to call it? Rootsmagic would firstly have to facilitate direct geocoding from an embedded map and also provide the option to enable reverse geocoding but the pin in the map in the most accurate definition of where the event took place and could be refined as information is uncovered.

 


"Never, for the sake of peace and quiet, deny your own experience or convictions"

— Dag Hammarskjold

 

Current user of Rootsmagic version 7.5.7.0, Family Tree Maker 2014 and Legacy 7.5 on Win 10

 

Excel to Gedcom conversion - simple getting started tutorials here

 

Root


#2 Renee Zamora

Renee Zamora

    Advanced Member

  • Support
  • PipPipPip
  • 7851 posts

Posted 12 March 2018 - 03:45 PM

Confirming this is on the enhancement request list. 


Renee
RootsMagic

#3 Jerry Bryan

Jerry Bryan

    Advanced Member

  • Members
  • PipPipPip
  • 3024 posts

Posted 28 April 2018 - 04:12 PM

So maybe we have been doing it all wrong, maybe in a lot of cases we should be sticking a pin in the map and not endlessly deliberating over what to call it? Rootsmagic would firstly have to facilitate direct geocoding from an embedded map and also provide the option to enable reverse geocoding but the pin in the map in the most accurate definition of where the event took place and could be refined as information is uncovered.

 

I had been meaning for a long time to get back to this message, and never got around to it.

 

The context is that it is very difficult to interchange genealogical data about place names. One issue is simply "genealogical standards" writ large. For example, do you call the country of which I am a citizen the United States of America (official name), United States, America, USA, U.S.A., US, U.S., etc. Different large and influential web sites do it differently. Various influential experts recommend differently. And that doesn't even get into support for multiple languages. One well known and influential language calls my country États Unis, and the examples abound. Germany is Deutchland and Tyskland and Allemagne and many other names depending on the language you are using.

 

And time frames are important. One of my families that I research mostly heavily immigrated from Germany, but should I call it Germany or should I call it Bavaria since there was no Germany at the time of their immigration in the 1700's? Some experts say to record the place name as of the time of the event so you can find the records and other experts say to record the modern place name so it can be geocoded.

 

What I have been musing about for several years has been the need for some sort of numeric coding of Place Names for data interchange and that Germany and Deutchland and Tyskland and Allemagne would all use the same numeric code that would be rendered by software into the appropriate language for the user. The user should never have to see or know about the numeric code that's behind the scenes. And separating out the behind the scenes coding of place names and the way place names are interchanged on the one hand from the way place names appear in reports on the other hand might quell this constant debate about how best to enter place names.

 

I suppose that begs the question of what the numeric code would be for Bavaria since there is not a one-to-one correspondence between the concept of "Bavaria" and.the concept of "Germany". And indeed, all kinds of borders and all kinds of Place Names change all the time. Be that as it may, I think the numeric coding for data storage and data interchange would be a step in the right direction.

 

I think that Vyger is proposing something similar to my numeric codes except that the "unique number" that would be used behind the scenes for country codes (and I suppose for for state codes and county codes and city codes and codes for any other kind of place) would be a latitude/longitude coordinate, i.e, a pin on a map. I guess I have a hard time seeing how that would work. For one thing, where would the pin go for the United States of America? A logical place might be somewhere in Kansas since it's sort of in the middle (well, it's in the middle if you ignore Alaska and Hawaii). Another logical place might be Washington, D.C. because it's the capital. If standards were to go this route, I think instead of a pin there would have to be a GPS polygon outlining the place. But the GPS polygon would have to vary in time.

 

The "vary in time" problem seems to wreak havoc with almost any system that otherwise seems reasonable. Here's an example. A big focus of my research is Rockingham County, Virginia. Many of my ancestral lines lived there for a generation or two in the 1700's before moving to Tennessee in the late 1700's or sometimes in the early 1800's. I'm used to searching records for my ancestors in Augusta County, Virginia before 1778 because Augusta County was split to form Rockingham County in 1778. But I recently found an important deed for my ancestors in Rockingham County that was registered in Orange County, Virginia in 1749. That seems very strange when you consider that Augusta County was formed from Orange County in 1738 and that therefore a deed in what is now Rockingham County should have been registered in Augusta County in 1749. But Augusta County was not really organized until 1745 and many deeds continued to be registered in Orange County for several years after that. I have managed to create an accurate GPS polygon for this deed which I can feed to Google Maps and place the parcel of land correctly in modern day Rockingham County. So why would I want a pin for Orange County to be associated with this deed?

 

Jerry



#4 Vyger

Vyger

    Advanced Member

  • Members
  • PipPipPip
  • 3144 posts

Posted 29 April 2018 - 09:48 AM

My proposal for reverse geocoding was born out of a previous post of your on Country Codes on another thread. Any such Country Code would never be agreed, one provider would try to better another and as we know borders change over time so latitude and longtitude are the perfect solution, the pin in the map.

 

I think that Vyger is proposing something similar to my numeric codes except that the "unique number" that would be used behind the scenes for country codes (and I suppose for for state codes and county codes and city codes and codes for any other kind of place) would be a latitude/longitude coordinate, i.e, a pin on a map. I guess I have a hard time seeing how that would work. For one thing, where would the pin go for the United States of America? A logical place might be somewhere in Kansas since it's sort of in the middle (well, it's in the middle if you ignore Alaska and Hawaii). Another logical place might be Washington, D.C. because it's the capital. If standards were to go this route, I think instead of a pin there would have to be a GPS polygon outlining the place. But the GPS polygon would have to vary in time.

 

My proposed solution to this very valid point is a few additional small fields in the place table, one indicating whether the place was Auto geocoded by Rootsmagic or Manually geocoded by the user. This now leaves 3 variants Auto, Manual and Null all of which should be filterable for further research.

 

I would further qualify a manually geocoded Place with a selection for Exact or Approximate, if Approximate was chosen then allow a follow on block selection for radius of approximation. None of these additions would hamper the user who just wants to record "any old place" in their own personal style, the option to progress to geocoding should be selectable the those "any old places" will just exist with a NULL flag in the Place Table for anyone who cares enough later to resolve them.

Personally I would never geocode a Country or Continent as it seems pointless but for those in other parts of the world where the text notation is illegible to them there is some value in having this Auto geocode, the proposed Auto flag would indicate it as such.

 

The "vary in time" problem seems to wreak havoc with almost any system that otherwise seems reasonable. Here's an example. A big focus of my research is Rockingham County, Virginia. Many of my ancestral lines lived there for a generation or two in the 1700's before moving to Tennessee in the late 1700's or sometimes in the early 1800's. I'm used to searching records for my ancestors in Augusta County, Virginia before 1778 because Augusta County was split to form Rockingham County in 1778. But I recently found an important deed for my ancestors in Rockingham County that was registered in Orange County, Virginia in 1749. That seems very strange when you consider that Augusta County was formed from Orange County in 1738 and that therefore a deed in what is now Rockingham County should have been registered in Augusta County in 1749. But Augusta County was not really organized until 1745 and many deeds continued to be registered in Orange County for several years after that. I have managed to create an accurate GPS polygon for this deed which I can feed to Google Maps and place the parcel of land correctly in modern day Rockingham County. So why would I want a pin for Orange County to be associated with this deed?

 

Firstly, I never record Places as some indication of where records may be held, I only record the factual Place the event took place. Knowing the administrative changes over time and were records may or may not be available and the other possible repositories and registries surrounding those areas is a different knowledge for the researcher.

 

Sadly the reference material to support "vary in time" is not totally available at this time, at least to my knowledge, but the database supporting County Check is a beginning if a financial hook can be identified to a clever developer in the future.

 

As a footnote, I have seen most variants of Place recording over my years, good and bad, I have engaged in numerous discussions on the subject and noted how personal preference for the written word have in many cases continued to prevail over any sort of standard. I have seen those attempts at standards picked apart and critized on numerous occassions to the point where I have all but given up discussing the topic. Through all those threads, opinions and discussions a standard which is accepted world wide already exists, it traverses language, locally known as, personal styles etc. we use it everyday but not always in our genealogy and its called GPS.


"Never, for the sake of peace and quiet, deny your own experience or convictions"

— Dag Hammarskjold

 

Current user of Rootsmagic version 7.5.7.0, Family Tree Maker 2014 and Legacy 7.5 on Win 10

 

Excel to Gedcom conversion - simple getting started tutorials here

 

Root