Concerning responsible use and publication of crisis data for South Sudan


This blogpost seeks to answer a number of honestly expressed concerns – which have been recently conveyed to us – about possible risks attached to the publication of different kinds of information from the public domain (principally Twitter names, radio stations and event locations) in a data resource doc for South Sudan that Sara-Jayne Terp of OpenCrisis set up before Christmas 2013 and which has since been curated by myself, Brendan O’Hanrahan.

We have examined these concerns in great detail and have sought and received the views of several humanitarian information and media professionals who are either based in South Sudan or were based there until the outbreak of this crisis,

Twitter Lists

We believe that there is a fundamental misapprehension at the heart of the concerns that have been raised about our work, particularly as it relates to Twitter: contrary to those assumptions, South Sudan is not experiencing a (so-called) “Arab Spring” or Iranian “Green Revolution” situation at the minute – rather, it is a minor civil war with ethnic undertones, which is essentially based on two factions jockeying for control of one party at the heart of what is, effectively, a one-party state. I wouldn’t classify anybody on the Twitter list contained within our resource doc as a proper social media activist in the sense one would use this term for other recent political crises: there’s one (very high-profile) rap artist in the diaspora who has strongly identified with one side, but somebody who appears on the BBC, Al Jazeera, etc doesn’t really need his identity protected. Then there are 3-4 people who might be categorised as peace activists – but while in some countries this might leave them vulnerable to persecution of some kind, there has been absolutely zero evidence of this for South Sudan. Indeed, if anything, my analysis would be that the peace activists have been more aligned with the status quo than against it – with more of their efforts focused on getting the rebels to abandon their rebellion…and being critical of the UN in a way that sometimes echoes regime criticism.

Twitter does not exist in some grey samizdat-style space; thanks to Twitter lists (not to mention blogposts like Lesley Anne Warner’s), anybody who wants to get hold of prominent Twitter accounts relating to South Sudan can do so in seconds.  To illustrate, here’s a table of the most well-known Twitter lists for (or including) South Sudan.

List Name No. of a/cs Owner
Sudan/South Sudan 69 Gintare Eidimtaite
South Sudan 12 Quinn MacDonald
South Sudan 19 Rachael Akidi
South Sudan 39 Alun McDonald
South Sudan 26 Veronica Ledesma
South Sudan 72 Gabriela Jacomella
South Sudan 87 Geoffrey York
South Sudan 273 Brendan O’Hanrahan
South Sudan 24 Daniel Stuckey
southsudan 81 Clemens Lerche
SS 11 AbdulRahman
Sudans 90 Robyn Dixon
South Sudan 27 Paul Brown
Sudans 125 Lauren P Blanchard
SouthSudanCrisis 100 Carlo Pisani
Sudanese 169 John Fenning
South Sudan 47 Ben Byerly
R.O.S.S. Related 100 Christine Elizabeth
south sudan 55 Hosanna Fox
South Sudan 75 U.S. Mission in Juba
South Sudan 30 Akshaya Kumar
Sudan/South Sudan 164 Teresa Reiter
SouthSudan 41 Anna Bagadi
both sudans 142 sonjasugira

And, in the interests of shining more light about what sort of people actually populated the resource doc’s Twitter list, I extracted the following summary breakdown

South Sudanese citizens No. in (or mainly in) South Sudan Diaspora Journalists/news outlets Analysts Humanitarian or UN
21 12 11 23 11 4

One minor point to highlight here is that the resource doc’s Twitter page has very few full names (this is simply because it was abandoned as a continuing project – because it was felt that public Twitter lists did the same job pretty much as adequately) – and those few nearly all relate to journalists and analysts. Though considering how conspicuous anybody on that list is, I really don’t see a problem with putting their full names there. It would save some evil mukhabarat-style agent 15 minutes of due diligence at most. A media and social media specialist normally based in South Sudan and with extensive experience of social media dynamics in East Africa concurred with our view that there were no vulnerable activists on this list.

Radio Station Lists

I am aware that radio stations (as well as several other kinds of media outlets) have been targeted in South Sudan – there is absolutely no question that South Sudan has become an increasingly hostile environment for journalists to work in over the last 18 months – see the many examples detailed on CPJ’s South Sudan page. However, again, I really, really doubt whether any of the information provided in our document would be necessary for any media suppression goons to do their job. From the government side all radio stations are licensed – so they have that information already…and from the rebel side, it does not make a plausible scenario that they would use our information to potentially target any radio operations when they have local knowledge already at hand. It should also be mentioned that there is already a detailed map of radio stations for South Sudan.

It’s important to stress that we deliberately did not keep any notes as to stations’ political alignment nor are any details about names of those working for them or their actual addresses given.

We have spoken to professionals working in the media development and humanitarian information spheres for South Sudan (two present in-country) – showing them the resource doc, and they have said that as long as no names or detailed addresses were published they saw no problems, since such information is widely available, or the radio stations are well known in their local areas.

Placenames and associated information

The source of over 50% of the named locations (and all of the IDP locations) in this doc is the body of humanitarian updates issued, principally, by OCHA, as well as IOM and USAID. See e.g., this and this, so any condemnation of this ‘loose’ information policy is effectively extended to them. 95% of the remainder of the locations in the placenames and coordinates sheet originate in news reports, or from Twitter updates from trusted journalists  – and the vast majority appeared in several different news sources. Many of the placenames first mentioned in news reports subsequently appeared in the maps of violent incidents produced by USAID, e.g. this one for 27 January. The remainder are either from otherwise researched and credible sources or are included for assistance in geographical orientation.

While there is no direct verification protocol, only reports from trusted sources have been included  (and since I have been following this crisis from close to the beginning in daily detail and been in dialogue with journalists and analysts who know the situation intimately, I think I am relatively qualified to make judgements as to who is trusted and who isn’t) – and if there was still some doubt attached to them, this is mentioned in the notes column. There were no attempts to categorise or otherwise analyse the events mentioned. This is because while it was thought that there would be organised attempts to map the violence and the humanitarian crisis, in practice the only non-OCHA, IOM, Enough Project or USAID mapping initiative was an abortive one, which was suspended after 2-3 days close to Christmas.  The purpose of the placename map therefore evolved to be a resource for either any media people who might seek to know where particular places were, or as the basis for assisting humanitarian mapping projects. One example of this were the maps I produced to help HOT in their identification of some of the priority areas for their South Sudan volunteer mapping response (in support of UNITAR/UNOSAT). It was never, and this is important to stress, intended as the basis for an event map or a violence map – the notes on what put that place “in the news” were just intended for context.

Which brings us to the specific claims that the mentioning of these places brought inherent danger with it to any people staying at those locations. The claim about people being put in danger because the location of a refugee camp is given doesn’t really withstand proper scrutiny – these refugee camps are BIG.  All of them with the exception of those at Minkamman and the UNMISS compounds are existing entities which are extensively documented (and mapped – see especially UNITAR’s collection of maps for South Sudan here) all over the Internet,  and it simply wouldn’t be possible to come into an area and talk to local people without becoming aware of them. The new ones have also been extensively publicised and mapped elsewhere (again, especially via UNITAR). Neither does it make any sense that mentioning e.g., a village which has been raided would make that village vulnerable to another attack in some way.

We spoke to a prominent humanitarian information officer in South Sudan and they said that they had seen no problem in mapping IDP locations and expressed no concern about mapping other locations included in the database.

We have honestly tried to conceive of (realistic!) scenarios whereby some bad actors would use this document to plan or inform their malfeasance, but we honestly couldn’t find any that belonged in the realm of commonsense.

Circulation

We have only recently put this up on the OpenCrisis site (although frankly you wouldn’t find it unless you knew to look for it), and the version there is published without any notes on the events associated with the placenames (though this is mainly because the metadocumentation of my agonising about where difficult-to-find places might be was considered to be probably rather tedious to casual visitors) and apart from the mention on Twitter of this version of the doc, it has otherwise had no circulation except to those involved directly in crisis mapping or media coverage of the crisis (e.g., OCHA in South Sudan, BBC, a couple of freelance journalists, MapAction, HOT). But, again we in OpenCrisis would like to stress that we see no a priori reason why individuals with legitimate interest in events in South Sudan should have less right to this information than those with organisational credentials. We believe in open data – while absolutely taking all due care that no harm is done – and by making these data available we hope to make a contribution, however small, to the deeper understanding and analysis of the complex recent events in South Sudan.

Which brings us on to the final and most important point – whether we have been sloppy in observing humanitarian criteria and being mindful of the safety of those who may be affected by the information in our document.

As the guidelines on the first page of the resource doc state – the overriding principle is to “do no harm”; and although it originally said “don’t identify people in-country,” it was felt that this was better and more practically phrased as “don’t give any details about people which may result in they (or their relations) being put at risk,” while maintaining the crucial first principle. Thus, we are extremely confident that nobody in any danger has been identified, i.e., all those named here are prominent on Twitter already (two possible exceptions are completely apolitical) and no personal information has been attached to either mappable data or radio station details. And, while verification of events associated with placenames has been limited to ensuring that only reports from trusted sources (e.g., OCHA, USAID, Radio Tamazuj, Gurtong Trust etc.) are noted, we feel that this has been to an appropriate level in view of the degree of risk attached to the publication of these names. We should have had a note explaining this for the document before, but we have now rectified this by adding such a clarification to the first page.

So, while acknowledging that it will not always be easy to reconcile the twin aims of making useful crisis data widely available and ensuring that no risk is thereby entailed, we honestly think that, at least for this project, we have managed to achieve a sensible balance.