Re: [hxlproject] Aid Worker Security Database HXL Tags

[David Megginson]

Hi from Ottawa, Meriah-Jo!  I will look at your proposed tags this week and send feedback.

Insecurity Insight collects data in a similar field, and have HXLated a number of their datasets, so you might find some inspiration there: https://data.humdata.org/search?ext_hxl=1&organization=insecurity-insight

Now that ACLED has moved behind a paywall, initiatives like AWSD and Insecurity Insight have become doubly important.

Cheers, David

--

David MEGGINSON (he/him)

Standards Lead, Centre for Humanitarian Data, UNOCHA

Mobile: +1-613-791-8553

Skype: david.megginson

TZ: Eastern North America (EST, UTC-0500; EDT, UTC-0400)

On Wed, 25 Aug 2021 at 11:25, Meriah-Jo Breckenridge <mjbr...@ualberta.ca> wrote:

Hello all,

I manage the Aid Worker Security Database (AWSD), a project of Humanitarian Outcomes funded by USAID, and I am new to the HXL group. As a proponent of open data, I commend the HXL tag project and think it is a great idea. I am in the process of adding HXL tags to our public dataset and thought it would be helpful to ask for any feedback from those who are familiar with HXL before we add the tags online. Attached is a sheet of 3 rows: 1) our column headings 2) my draft HXL tags 3) a sample incident. Please feel free to edit/suggest! Thank you in advance for your assistance.

For those that would like a little background on the AWSD; beginning in 2005 and published online in 2010, the AWSD is the only public source of aid worker attack data globally going back to 1997. We have over 120 INGO/NGO partners who share data confidentially annually, and we have public data-sharing partnerships with groups like ACLED and SNHR. All our anonymized incidents are available for download anytime on our website or through our API. We only collect data on serious aid worker attacks that we define as; killings, injuries requiring medical attention, kidnappings over 24 hours and rape/serious sexual assaults. For more info on our parameters: https://aidworkersecurity.org/about

Our team at Humanitarian Outcomes is very keen on sharing data across the humanitarian sector, so please let me know if anyone would like to chat further about aid worker security information as we are always open to partnerships.

Sincerely,

Meriah-Jo Breckenridge

Database Manager

Research Analyst

Humanitarian Outcomes

m:+1.780.937.6725

s: live:meriah-jo

    

--
You received this message because you are subscribed to the Google Groups "Humanitarian Exchange Language (HXL)" group.
To unsubscribe from this group and stop receiving emails from it, send an email to hxlproject+...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/hxlproject/61db5332-e093-49d4-a243-c505a281c604n%40googlegroups.com.

[Meriah-Jo Breckenridge]

Hi David, thank you very much for the quick reply and for looking at the tags! Really appreciate it. I forgot to mention Insecurity Insight is also a data-sharing partner of us and I used their set as an example when completing my own tags. 

Also, Ottawa! So rare I run into another Canadian. Hi from Edmonton :)

MJ

[Emerson Rocha]

This type of HXLated conventions related to targeted individuals, potentially with even greater level of details, is a common interest by regional initiatives!!!

Citing as example Brazil, there is this project https://nev.prp.usp.br/mapa-dos-grupos-armados-do-rio-de-janeiro/ and https://fogocruzado.org.br/ (this one with API, but not HXLated). I did not try, but very likely a few other countries in Latin America may have something like this.

One very, very desired feature I know from these projects is both some data standard to share at global scale and (even if obviously serious review) some place they could publish additional versions of the data like HDX. So @Meriah, if after reviewing with David HXL tagging conventions about this data, if in the next weeks (or months) someone could publish more about this use case (maybe a blog post, or some GitHub repository with tagging examples) this would be fantastic! Or, if for some reason even simpler conventions on this type of data could take too long to document, maybe if someone from AWSD, in addition to me, could kindly be in contact with some researchers on this topic in Brazil, I think we could help document this better.

One somewhat peculiar case about Brazil (since already have a lot of interest in such subjects and existing open source community) I believe if (after some data conventions) do exist some "open ideas" about what software could be done, very likely regional software developers already related to these projects would feel nudged (https://en.wikipedia.org/wiki/Nudge_theory) to create whatever new tools generic enough to be reused no just at their country. But before going to this point, they may be interested in at least being able to make proofs of concepts of converting different sources in a somewhat common HXLated format. Then if not ACLED, one or more contacts from AWSD on next months could help to give more confidence on such data conventions.

@David I was not aware that ACLED has more restrictions. It would be the reference I would use on discussions with Brazilian initiatives, so AWSD announced by Meriah as soon as it started to have HXLated content would be the new reference example.

Att.
Rocha

--

Emerson Rocha
Full stack developer at Alligo

Transdisciplinary researcher at Etica.AI

Member of  The IEEE Special Interest Group on Humanitarian Technology (IEEE SIGHT)

Member of The IEEE Global Initiative on Ethics of Autonomous and Intelligent Systems

Supporter (at multiple local sections/structures) of Amnesty International

[David Megginson]

Thanks again for sharing that, Meriah-Jo. Since you were kind enough to share your proposed tags publicly here in the forum, I'll share my feedback that way as well, and give other members a chance to jump in.

You did a great job choosing the hashtags. There's nothing incorrect about the ones you used, so please take the following as suggested refinements:

• HXL has a core hashtag #group for a non-aid actor, so you could replace #x_actor_name and #x_actor_type with #group+name and #group+type
• Instead of #x_country_code, you can use #country+code
  
• If you refine #event to #event+id, it will give you the opportunity to also use #event+type (or #event+context) and #event+means instead of #x_attack_context and #x_attack_means
• The core hashtag #date is flexible, and you can use #date+year, #date+month+, and #date+day instead of #year, #month, and #day; it is also a good idea to add an extra column with the full date in ISO format (YYYY-MM-DD), since it makes it easier to sort the data, display it on a timeline, etc.
• HXL 1.1 changed the standard attributes for language to add "i_" (for "internationalisation") before them, to avoid any conflict with other attributes, so the attribute for English would be "+i_en" rather than "+en" now.
• You could optionally refine #loc to #loc+type, since you're giving the type of location rather than its name

Please feel free to get in touch (either privately or publicly here in the list) if you have any other questions, and I'll be happy to review a v2 if you'd like.

Cheers, David

--

David MEGGINSON (he/him)

Mobile: +1-613-791-8553

Skype: david.megginson

TZ: Eastern North America (EST, UTC-0500; EDT, UTC-0400)

[David Megginson]

Looks great! I'm looking forward to seeing the data on HDX (and other aid-data sites).

Cheers, David

--

David MEGGINSON (he/him)

Mobile: +1-613-791-8553

Skype: david.megginson

TZ: Eastern North America (EST, UTC-0500; EDT, UTC-0400)

On Thu, 2 Sept 2021 at 11:01, Meriah-Jo Breckenridge <mjbr...@ualberta.ca> wrote:

Dear David and Emerson,

Thank you both for taking a moment to review my first email. Emerson, thank you for linking those interesting groups in Brazil. Unfortunately, aid worker attack data is often very limited in central and south America so I am always interested to make contacts there. I will reach out to their group to see if any collaboration is possible.

David, thank you so much for all the edits/suggestions! I've attached version 2 if you would be so inclined to help again. I did have one question about our motive category, would #event+motive be an option?

Also for anyone else in the group, please let me know if anyone wants to connect on aid worker security data more generally.

Sincerely,

MJ

You received this message because you are subscribed to a topic in the Google Groups "Humanitarian Exchange Language (HXL)" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/hxlproject/pVkNgPKTZ0k/unsubscribe.
To unsubscribe from this group and all its topics, send an email to hxlproject+...@googlegroups.com.
To view this discussion on the web, visit https://groups.google.com/d/msgid/hxlproject/CABCr9BhCHpznhVPsFTOvQ%2BndBV57_2y9cx95Xr8hLvfZ28c_5A%40mail.gmail.com.

--

Meriah-Jo Breckenridge

Consultant

+1 780 937 6725

mjbr...@ualberta.ca

--
You received this message because you are subscribed to the Google Groups "Humanitarian Exchange Language (HXL)" group.
To unsubscribe from this group and stop receiving emails from it, send an email to hxlproject+...@googlegroups.com.

To view this discussion on the web, visit https://groups.google.com/d/msgid/hxlproject/CAPpwXaL9LriLQ8uOqMq%2Bp8yKW3_VEfgqsApdbwm5hdFnZ0ggTQ%40mail.gmail.com.