Need API for events and projects

Started by Private User on Tuesday, June 14, 2011

Participants:

Related Projects:

Showing all 11 posts
Private User
6/14/2011 at 7:43 AM

There is a pretty interesting effort to document the Civil War in geni. The discussion for it is here:
http://www.geni.com/discussions/93453?msg=711345

I'm trying to help by automating some of the work but there are a couple of APIs that I need to make this happen:
- API to create an event, set the dates, location, etc
- API to retrieve and edit the event's "About" text
- API to add a profile to an event (not critical but would be handy)
- API to create a project
- API to retrieve and edit the project's "About" text

Scott Steadman and Noah Tutak any chance you guys could add this? The numbers involved for this civil war documentation effort are huge (6000 companies, 3 million soldiers, 20000 battles). It needs some automation help.

Thanks

Private User
6/14/2011 at 8:35 AM

Let me add that from what I'm gathering - the end result would be unique and not available ANYWHERE else - print or internet.

We have a Civil War database / website builder who can confirm this, I think - John Rigdon - he is making his collected Civil War data available to "merge" with Geni profiles.

George Gearhart I am smelling the marketing opportunity big time.

6/14/2011 at 8:42 AM

I second the need to be able to have an API to retrieve and edit the project's "About" text. If we had it I'd be using it on the http://www.geni.com/projects/Namesakes and http://www.geni.com/projects/Original-Proprietors-Hartford-CT projects right now.

6/14/2011 at 9:18 AM

Yes, it would be unique. I can comment on the various sites which have implemented various parts, but noone comes close - including myself - to having a comprehensive site.

You folks at Geni have a unique set of individual data which in my opinion is worlds ahead of anyone else. In my opinion this Civil War Project could become the focal point, past and future, for researchers to tie things together between their family trees and the historical / socialogical universe in which they lived.

Private User
6/14/2011 at 9:23 AM

John Rigdon wrote:

===
<this project> could become the focal point ... for researchers to tie things together between their family trees and the historical / socialogical universe in which they lived.

===

That is exactly what has me so interested in Projects. It takes family research to the next level by adding historical, cultural, geographic, sociological, linguistic and anthropological layers.

Face it - long lists of "begats" are not interesting. Even visiting wonderful places like London becomes so much more vital when you start to realize, "my ancestor walked these streets."

I really believe geni is ahead of the curve already, but needs to solidify and innovate ... not imitate! more. If all it takes is an API? Oh, please, just do it. :)

6/14/2011 at 10:07 AM

This is definitely interesting, and for far more reasons that just marketing.

There are some considerations that we'll have to discuss internally. Some of them:

1. We plan to create a historical events feature (and I don't know how similar to or different from a project it would be). It's not on the immediate horizon, but I'll ask about it today and hopefully get a better idea for when we roll it out.

2. If/when we do add a historical events feature, we will likely capture more meta data about the project than just an About wikitext. Things like the date(s) on which the event occurred, the place(s) where the event happened, etc. So we'd need to be able to figure out how to migrate a project to a historical event without it getting too messy.

3. What is the best way to handle groups of people when tagging them to an event or project? (Examples: an army, a company within the army, government bodies)

4. There is a lot to figure out with events in general. The Civil War is a great example of the complexity: The Civil War, in and of itself, is an event. However, so are every battle, every political event (the secession of the Confederacy, the Emancipation Proclamation, the Gettysburg Address, Lincoln's assassination), and probably a lot of other things that I'm not thinking of.

A completely different example: Woodstock was an event, but so was the Jimi Hendrix set at Woodstock.

For many types of events, there is a clear parent-child relationship (Jimi wouldn't have played a set at Woodstock if Woodstock didn't happen, and the Battle of Antietam wouldn't have occurred without a Civil War). But for many event types, there is no clear hierarchy like this.

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

When we release events, it is more likely that we will release them in a manner that can apply to all of history. If projects are a suitable means of handling events in the interim, than it is probably worth spending a little time making sure that the projects will easily transition to events.

I will begin a conversation internally about this. It's exciting to see everyone's enthusiasm about it :)

Private User
6/14/2011 at 11:02 AM

Cool, it sounds like you guys have some stuff in mind already. Ok so assuming geni implements historical events here are some thoughts

- need a way to convert an event created for a profile to an 'historical event'. This way we could go ahead and create an event for the Battle of Gettysburg today and convert it down the road. Or we could just wait until historical events are out.

- keep using projects to create groups of profiles

- a historical event needs to be able to add a project as a participant

- once a project is added to a historical event all of the profiles in that project should have that event on their timeline with an option to remove the profile from the event. One of the scenarios this is needed for is when a soldier was a POW and did not participate in some battles that his company was in.

- a project can have multiple parent projects

- a project's relationship to a parent project needs an option to be bound to ranges of dates

Here is one example John sent me of where a project (GA 49th infantry) would need multiple time dependent parents projects:
GA 49th Infantry Regiment

GEORGIA FORTY-NINTH INFANTRY REGIMENT
Army of Northern Virginia
Confederate States of America

Assignments

J. R. Anderson's Brigade, Department of North Carolina (April-May 1862)

J. R. Anderson's Brigade, Department of Northern Virginia (May 1862)

J. R. Anderson's Brigade, A. P. Hill's Division, Army of Northern Virginia
(May-June l862)

J. R. Anderson's-Thomas Brigade, A. P. Hill's Division, 1st Corps, Army of
Northern Virginia (June-July 1862)

Thomas' Brigade, A. P. Hill's Division, 2nd Corps, Army of Northern
Virginia (July 1862-May 1863)

Thomas' Brigade, Pender's-Wilcox's Division, 3rd Corps, Army of Northern
Virginia (May 1863-April 1865)

If you guys pulled this off that would be one heck of an infrastructure for complex events like wars.

6/14/2011 at 11:05 AM

The historical events idea plays to the heart of one of the reasons I found Geni so very interesting. As Erica can attest, I love timelines. Do keep us posted how this development comes along...

6/14/2011 at 7:32 PM

Daniel,

I think you see the complexity :) We met briefly today to discuss this and places pages, and I think we'll probably tackle the initial version of places first. There is an awful lot that we have to consider for historical events, and it's likely going to be months before we can roll it out due to the sheer amount of work.

That said, there's a decent chance that an event will be similar in many ways to a project (as will a place page), so projects could be a good way to get started. Ideally we'd figure out some markup that would make it easy to parse relationships of all of these projects for future migration purposes.

Regarding more project calls in the API, I can think of some spam-related reasons that we would want to be careful, but I will investigate more this week.

-George

Private User
6/15/2011 at 5:42 AM

Yes I can see how once you create a system for time dependent groups of people (projects), time dependent places, and time dependent events that can all have multiple relationships/dependencies with each other you would have something very flexible yet so complex nobody could use it. Of course you'd also have a platform that probably no other company has so it might be worth it if you can make the GUI manageable.

6/15/2011 at 9:49 AM

Bingo :)

Showing all 11 posts

Create a free account or login to participate in this discussion