All people in one view: a non-traditional visualisation option

classic Classic list List threaded Threaded
4 messages Options
Reply | Threaded
Open this post in threaded view
|

All people in one view: a non-traditional visualisation option

J M Wynands
Hi folks,

I read with interest the conversation(s) recently about interactive trees, & wanted to add another new way to visualise family history. It's fairly non-traditional but can produce a picture of all people in one view plus yield some interesting insights.

For especially large trees where the path back in time isn't as much of a concern an actor-network diagram can be useful. Actor-network diagrams are mainly used for social network analysis, eg of contemporary Facebook networks or the Enron emails. They're relatively simple, composed of just nodes (people) & edges (relationships), although you can add a range of attributes to those two parts to give more depth to the picture. These can include node shapes by type, eg gender (similar to a genotype graph), node colours by type, eg family or century, node sizes by edge count (more relationships = larger nodes), plus edges can be coloured & given greater thickness according to relationship type & strength.

Tools you can use to produce these diagrams include Polinode, which is cloud based, & Gephi, which has a desktop version that can output to html, if you need/want to publish the result. The tools use some pretty sophisticated algorithms to organise the network into clusters, according to which nodes have what edges (beyond me personally to understand, but if you're better at maths & stats than I, there's plenty to learn there).

The way actor-network diagrams generate clusters really helps with understanding how people have/are connected & over time family groups have formed & morphed. It also seriously helps make sense of large data sets. In my case I inherited my grandmother's 30+ years of research, which totals 11,451 entries. The majority are from the last 3 centuries only, so it's more a 'wide net' family history than a traditional genealogy anyway. It features among other things several intermarriages at different times, because Australia in the late 18th, 19th & early 20th centuries was really not that big a place.

Since it was all created in paf I'm eternally grateful for Gramps for providing a way forward with the data; I've also used the navweb output & that feature is much appreciated too. However I couldn't make much of all the unfamiliar names, either in Gramps' different trees, or especially among the mountains of paper that I was given to digitise (how do you create a logical system for organising & naming files when each document has a dozen different names present? The mind boggles). So I used Polinode, after trying Gephi, to produce an interactive map of all 11,451 in one view. The result is both a fantastic demonstration of the scale of my grandma's research, plus a way to comprehend the data & reams of paper: via the main clusters, which I've developed into 'clans'. If you'd like to see the map, please go to https://indidem.github.io/.

If you're open to trying a fairly novel, non-traditional visualisation, you may find that Polinode &/or Gephi add some value to your research. All people in one view is difficult to achieve with trees, & even interactive ones can leave you lost after a few jumps. Both programs have good communities (Polinode is newer than Gephi but the developer/owner is a fast & generous responder to forum queries, even for free account holders), plus there's tutorials on youtube & other help/blog resources to draw on to learn the ropes.

Cheers,
Jess


--
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users
https://gramps-project.org
Reply | Threaded
Open this post in threaded view
|

Re: All people in one view: a non-traditional visualisation option

StoltHD
I also use Network Graphs in my research...
But I am creating a knowledge graph with relations not only between people, but between repositories, sources, documents, places, events, people (both familiy- and "external" relations).

There are multiple software in additon to Polinode and Gehpi...
- yED Desktop, can import gedcom (Free)
- Cytoscape, Imports multiple formats, but not gedcom (open source, free)
- Zegrada (Open Source Free)
- VIS-Desktop (Open Source, free, Not recommended for large scale)
- Social Network Visualizer (Open Source, free)

- Palladio is another advanced analyzing tool, (open source, free) this also has GPS mapping, histograms and facet search, in addition to a gallery and table grid view...
Advanced concept map tools:
VUE, (Not maintained anymore, but still work)
cMap Tools (free)
Instrumind ThinkComposer (Windows only, free)
---------------------------

I n addition there is Neo4j, the graph database, it can connect to mongoDB and utilize the Gramps mongoDB database.

It actually would be possible to create this type of view in Gramps by making a add-on/gramplet/viw, Graphviz supports this type of networks out of box


At the moment I use OpenRefine to clean data (csv at the moment) and merge with reconcile with other datasets.


Because of the size and complexity of my project, I'm actually looking at solutions as Omeka, The Arches Project, Researchspace, Histograph and Heurist (an Australian project, http://heuristnetwork.org/ ).


---------------------------------------------------------------

I really liked your layout of the website, it was clean and easy to read and navigate... really good work there...
Could you share what other tools you used?

Jaran

søn. 8. des. 2019 kl. 00:13 skrev J M Wynands <[hidden email]>:
Hi folks,

I read with interest the conversation(s) recently about interactive trees, & wanted to add another new way to visualise family history. It's fairly non-traditional but can produce a picture of all people in one view plus yield some interesting insights.

For especially large trees where the path back in time isn't as much of a concern an actor-network diagram can be useful. Actor-network diagrams are mainly used for social network analysis, eg of contemporary Facebook networks or the Enron emails. They're relatively simple, composed of just nodes (people) & edges (relationships), although you can add a range of attributes to those two parts to give more depth to the picture. These can include node shapes by type, eg gender (similar to a genotype graph), node colours by type, eg family or century, node sizes by edge count (more relationships = larger nodes), plus edges can be coloured & given greater thickness according to relationship type & strength.

Tools you can use to produce these diagrams include Polinode, which is cloud based, & Gephi, which has a desktop version that can output to html, if you need/want to publish the result. The tools use some pretty sophisticated algorithms to organise the network into clusters, according to which nodes have what edges (beyond me personally to understand, but if you're better at maths & stats than I, there's plenty to learn there).

The way actor-network diagrams generate clusters really helps with understanding how people have/are connected & over time family groups have formed & morphed. It also seriously helps make sense of large data sets. In my case I inherited my grandmother's 30+ years of research, which totals 11,451 entries. The majority are from the last 3 centuries only, so it's more a 'wide net' family history than a traditional genealogy anyway. It features among other things several intermarriages at different times, because Australia in the late 18th, 19th & early 20th centuries was really not that big a place.

Since it was all created in paf I'm eternally grateful for Gramps for providing a way forward with the data; I've also used the navweb output & that feature is much appreciated too. However I couldn't make much of all the unfamiliar names, either in Gramps' different trees, or especially among the mountains of paper that I was given to digitise (how do you create a logical system for organising & naming files when each document has a dozen different names present? The mind boggles). So I used Polinode, after trying Gephi, to produce an interactive map of all 11,451 in one view. The result is both a fantastic demonstration of the scale of my grandma's research, plus a way to comprehend the data & reams of paper: via the main clusters, which I've developed into 'clans'. If you'd like to see the map, please go to https://indidem.github.io/.

If you're open to trying a fairly novel, non-traditional visualisation, you may find that Polinode &/or Gephi add some value to your research. All people in one view is difficult to achieve with trees, & even interactive ones can leave you lost after a few jumps. Both programs have good communities (Polinode is newer than Gephi but the developer/owner is a fast & generous responder to forum queries, even for free account holders), plus there's tutorials on youtube & other help/blog resources to draw on to learn the ropes.

Cheers,
Jess
--
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users
https://gramps-project.org


--
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users
https://gramps-project.org
Reply | Threaded
Open this post in threaded view
|

Re: All people in one view: a non-traditional visualisation option

Connie Mack
Great job, Jess. I'm impressed. Thanks for sharing.


On Sun, Dec 8, 2019 at 3:47 AM StoltHD <[hidden email]> wrote:
I also use Network Graphs in my research...
But I am creating a knowledge graph with relations not only between people, but between repositories, sources, documents, places, events, people (both familiy- and "external" relations).

There are multiple software in additon to Polinode and Gehpi...
- yED Desktop, can import gedcom (Free)
- Cytoscape, Imports multiple formats, but not gedcom (open source, free)
- Zegrada (Open Source Free)
- VIS-Desktop (Open Source, free, Not recommended for large scale)
- Social Network Visualizer (Open Source, free)

- Palladio is another advanced analyzing tool, (open source, free) this also has GPS mapping, histograms and facet search, in addition to a gallery and table grid view...
Advanced concept map tools:
VUE, (Not maintained anymore, but still work)
cMap Tools (free)
Instrumind ThinkComposer (Windows only, free)
---------------------------

I n addition there is Neo4j, the graph database, it can connect to mongoDB and utilize the Gramps mongoDB database.

It actually would be possible to create this type of view in Gramps by making a add-on/gramplet/viw, Graphviz supports this type of networks out of box


At the moment I use OpenRefine to clean data (csv at the moment) and merge with reconcile with other datasets.


Because of the size and complexity of my project, I'm actually looking at solutions as Omeka, The Arches Project, Researchspace, Histograph and Heurist (an Australian project, http://heuristnetwork.org/ ).


---------------------------------------------------------------

I really liked your layout of the website, it was clean and easy to read and navigate... really good work there...
Could you share what other tools you used?

Jaran

søn. 8. des. 2019 kl. 00:13 skrev J M Wynands <[hidden email]>:
Hi folks,

I read with interest the conversation(s) recently about interactive trees, & wanted to add another new way to visualise family history. It's fairly non-traditional but can produce a picture of all people in one view plus yield some interesting insights.

For especially large trees where the path back in time isn't as much of a concern an actor-network diagram can be useful. Actor-network diagrams are mainly used for social network analysis, eg of contemporary Facebook networks or the Enron emails. They're relatively simple, composed of just nodes (people) & edges (relationships), although you can add a range of attributes to those two parts to give more depth to the picture. These can include node shapes by type, eg gender (similar to a genotype graph), node colours by type, eg family or century, node sizes by edge count (more relationships = larger nodes), plus edges can be coloured & given greater thickness according to relationship type & strength.

Tools you can use to produce these diagrams include Polinode, which is cloud based, & Gephi, which has a desktop version that can output to html, if you need/want to publish the result. The tools use some pretty sophisticated algorithms to organise the network into clusters, according to which nodes have what edges (beyond me personally to understand, but if you're better at maths & stats than I, there's plenty to learn there).

The way actor-network diagrams generate clusters really helps with understanding how people have/are connected & over time family groups have formed & morphed. It also seriously helps make sense of large data sets. In my case I inherited my grandmother's 30+ years of research, which totals 11,451 entries. The majority are from the last 3 centuries only, so it's more a 'wide net' family history than a traditional genealogy anyway. It features among other things several intermarriages at different times, because Australia in the late 18th, 19th & early 20th centuries was really not that big a place.

Since it was all created in paf I'm eternally grateful for Gramps for providing a way forward with the data; I've also used the navweb output & that feature is much appreciated too. However I couldn't make much of all the unfamiliar names, either in Gramps' different trees, or especially among the mountains of paper that I was given to digitise (how do you create a logical system for organising & naming files when each document has a dozen different names present? The mind boggles). So I used Polinode, after trying Gephi, to produce an interactive map of all 11,451 in one view. The result is both a fantastic demonstration of the scale of my grandma's research, plus a way to comprehend the data & reams of paper: via the main clusters, which I've developed into 'clans'. If you'd like to see the map, please go to https://indidem.github.io/.

If you're open to trying a fairly novel, non-traditional visualisation, you may find that Polinode &/or Gephi add some value to your research. All people in one view is difficult to achieve with trees, & even interactive ones can leave you lost after a few jumps. Both programs have good communities (Polinode is newer than Gephi but the developer/owner is a fast & generous responder to forum queries, even for free account holders), plus there's tutorials on youtube & other help/blog resources to draw on to learn the ropes.

Cheers,
Jess
--
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users
https://gramps-project.org
--
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users
https://gramps-project.org


--
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users
https://gramps-project.org
Reply | Threaded
Open this post in threaded view
|

Re: All people in one view: a non-traditional visualisation option

J M Wynands
Thanks Connie :)
Jaran, fantastic list of resources there, thanks so much! I have plenty of data cleaning to do & may still need an archiving solution, regardless of using Wikitree. Last time I looked at DAM/CMS options about 4-5 years ago I found Musarch (musarch.com), Kora (github.com/matrix-msu/kora) & ResourceSpace (not researchspace); these others look more advanced & intriguing. 
Regarding my site, thanks for the praise :) It was a stretch for me, another new thing (new language & structure), but really it’s just a Jekyll site that uses a theme called Agency. The theme needed a fair bit of tweaking (still does) & is unsupported, but even so much of the hard work was done for me. 
Cheers & appreciate your interest,
Jess

On Mon, 9 Dec 2019 at 1:48 am, Connie Mack <[hidden email]> wrote:
Great job, Jess. I'm impressed. Thanks for sharing.


On Sun, Dec 8, 2019 at 3:47 AM StoltHD <[hidden email]> wrote:
I also use Network Graphs in my research...
But I am creating a knowledge graph with relations not only between people, but between repositories, sources, documents, places, events, people (both familiy- and "external" relations).

There are multiple software in additon to Polinode and Gehpi...
- yED Desktop, can import gedcom (Free)
- Cytoscape, Imports multiple formats, but not gedcom (open source, free)
- Zegrada (Open Source Free)
- VIS-Desktop (Open Source, free, Not recommended for large scale)
- Social Network Visualizer (Open Source, free)

- Palladio is another advanced analyzing tool, (open source, free) this also has GPS mapping, histograms and facet search, in addition to a gallery and table grid view...
Advanced concept map tools:
VUE, (Not maintained anymore, but still work)
cMap Tools (free)
Instrumind ThinkComposer (Windows only, free)
---------------------------

I n addition there is Neo4j, the graph database, it can connect to mongoDB and utilize the Gramps mongoDB database.

It actually would be possible to create this type of view in Gramps by making a add-on/gramplet/viw, Graphviz supports this type of networks out of box


At the moment I use OpenRefine to clean data (csv at the moment) and merge with reconcile with other datasets.


Because of the size and complexity of my project, I'm actually looking at solutions as Omeka, The Arches Project, Researchspace, Histograph and Heurist (an Australian project, http://heuristnetwork.org/ ).


---------------------------------------------------------------

I really liked your layout of the website, it was clean and easy to read and navigate... really good work there...
Could you share what other tools you used?

Jaran

søn. 8. des. 2019 kl. 00:13 skrev J M Wynands <[hidden email]>:
Hi folks,

I read with interest the conversation(s) recently about interactive trees, & wanted to add another new way to visualise family history. It's fairly non-traditional but can produce a picture of all people in one view plus yield some interesting insights.

For especially large trees where the path back in time isn't as much of a concern an actor-network diagram can be useful. Actor-network diagrams are mainly used for social network analysis, eg of contemporary Facebook networks or the Enron emails. They're relatively simple, composed of just nodes (people) & edges (relationships), although you can add a range of attributes to those two parts to give more depth to the picture. These can include node shapes by type, eg gender (similar to a genotype graph), node colours by type, eg family or century, node sizes by edge count (more relationships = larger nodes), plus edges can be coloured & given greater thickness according to relationship type & strength.

Tools you can use to produce these diagrams include Polinode, which is cloud based, & Gephi, which has a desktop version that can output to html, if you need/want to publish the result. The tools use some pretty sophisticated algorithms to organise the network into clusters, according to which nodes have what edges (beyond me personally to understand, but if you're better at maths & stats than I, there's plenty to learn there).

The way actor-network diagrams generate clusters really helps with understanding how people have/are connected & over time family groups have formed & morphed. It also seriously helps make sense of large data sets. In my case I inherited my grandmother's 30+ years of research, which totals 11,451 entries. The majority are from the last 3 centuries only, so it's more a 'wide net' family history than a traditional genealogy anyway. It features among other things several intermarriages at different times, because Australia in the late 18th, 19th & early 20th centuries was really not that big a place.

Since it was all created in paf I'm eternally grateful for Gramps for providing a way forward with the data; I've also used the navweb output & that feature is much appreciated too. However I couldn't make much of all the unfamiliar names, either in Gramps' different trees, or especially among the mountains of paper that I was given to digitise (how do you create a logical system for organising & naming files when each document has a dozen different names present? The mind boggles). So I used Polinode, after trying Gephi, to produce an interactive map of all 11,451 in one view. The result is both a fantastic demonstration of the scale of my grandma's research, plus a way to comprehend the data & reams of paper: via the main clusters, which I've developed into 'clans'. If you'd like to see the map, please go to https://indidem.github.io/.

If you're open to trying a fairly novel, non-traditional visualisation, you may find that Polinode &/or Gephi add some value to your research. All people in one view is difficult to achieve with trees, & even interactive ones can leave you lost after a few jumps. Both programs have good communities (Polinode is newer than Gephi but the developer/owner is a fast & generous responder to forum queries, even for free account holders), plus there's tutorials on youtube & other help/blog resources to draw on to learn the ropes.

Cheers,
Jess
--
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users
https://gramps-project.org
--
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users
https://gramps-project.org
--
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users
https://gramps-project.org


--
Gramps-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/gramps-users
https://gramps-project.org