[Qgis-user] Joining with Multiple Locations per Person

[Jake Maier]

Person A is linked to New York City, Person B is linked to New York City,
therefore Person A is linked to Person B.

 

Yes, however Person A and B also have 50 other locations that need to be
checked against each other. The data might say NY, NYC, or NY City or New
York, NY and other variations, geocoding allows for the different ways a
locations is represented I do not want to rely on how others spell their
locations and I need to share them on a map.

 

Isn't that relatively easy to fix by creating a new field, and assigning the
correct attribute by location, namely nyc or whatever name you choose?

Wouldn't that make things much easier

 

Jake

 

From: qgis-user-bounces at lists.osgeo.org
[mailto:qgis-user-bounces at lists.osgeo.org] On Behalf Of Lee Hachadoorian
Sent: Monday, November 18, 2013 1:10 AM
To: Sasa Sullivan; qgis-user at lists.osgeo.org
Subject: Re: [Qgis-user] Joining with Multiple Locations per Person

 

 

On 11/17/2013 03:34 PM, Sasa Sullivan wrote:

Hi Lee, 

 

Hopefully I can be a bit clearer here, if not please do not hesitate to let
me know, either on this post or to my email, I have tried to search online
and the help functions for similar questions and responses before I posted
to this board. Your questions/statements in bold text.

 

But I could be misunderstanding your intention. Some more detail would be
helpful.

 

The people concerned have been matched to each other via DNA, their mission
is to find the most recent common ancestors, this by looking at genealogical
data which consists of surnames and locations. I don't want to include the
questions about DNA as that is another matter and once I can find answers to
this question I can probably answer that one myself.

 

What is a record in you data?

 

A record consists of one person with multiple geographical locations and
surnames who need to find a common location(s)/surname(s) with other people
who also have multiple geographical locations and surnames (some small
groups of people are linked to each other as well). There is never just one
surname and/or location per person


This seems to be outside the "standard" GIS model where an entity has one
location (really one geometry) and a fixed number of clearly defined
attribute columns. Your entities have a variable number of locations, and
perhaps a variable number of attribute columns (the surnames) as well.
Possibly this has some similarity to an animal tracking model, where each
animal has multiple timepoints, and you might ask where two animals crossed
paths. But I am not very familiar with this field and don't know how easy it
would be to implement in QGIS or any other tool.




 

What location data do you have per person?

 

Each person's location numbers are varied, from four minimum to close to a
hundred

 

At a guess, this is something that can be done without any spatial data,
i.e.

 

Geocoding will work best as it does not rely on spellings with are too
varied

 

Person A is linked to New York City, Person B is linked to New York City,
therefore Person A is linked to Person B.

 

Yes, however Person A and B also have 50 other locations that need to be
checked against each other. The data might say NY, NYC, or NY City or New
York, NY and other variations, geocoding allows for the different ways a
locations is represented I do not want to rely on how others spell their
locations and I need to share them on a map.

 

Initially the people are linked to each other on one or more chromosomes, I
can sort that in Excel. There might be another type of programs that can
handle my questions however I do not know of any names, then I can transfer
the results into QGIS.

 

This is something I would tend to do in pure SQL.

 

I do not know much about programming, beyond what someone directs me to do
and was hoping the query functions of QGIS would handle what I was asking,
if this were a one to one comparison I could do it in Excel or a database
program, ultimately I need to put the information on a map

 

Thank you so much for responding, I don't know of a beginners list to ask
these questions. As I write this perhaps I can geocode and let Excel query
the latitude and longitude for my matches.


Possibly the geocoder outputs some other identifying field like a
standardized name, code, or some unique ID. In which case you could search
for matches on that field, rather than lat-long. 

One thing that might make this easier is to think of a
person_ancestry_locations table with a person ID and a location
(standardized name or code) field, where each person has one record for each
location they are associated with. 

But I can't think of a way to do this in QGIS that isn't incredibly
contorted. I would probably just use QGIS for the final mapping of common
locations.

Best,
--Lee 

-- 
Lee Hachadoorian
Assistant Professor in Geography, Dartmouth College
http://freecity.commons.gc.cuny.edu
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/qgis-user/attachments/20131118/a6b380c9/attachment.html>