These shapefiles were developed by volunteers of the Syria Tracker crowd project to support a better analysis of violent deaths in Syria resulting from the actions of the Assad Regime.
Note that Arabic fields are not equally supported in all GIS products-- unicode support is needed in order to view them. We've noticed that Apple-based platforms tend not to connect Arabic letters.
We apologize for the current lack of a formal data dictionary, but it has been a challenge to keep current. This will hopefully be enough to get you started.
We will summarize the files first, and then try to describe the attributes. The three files we will try to keep current are:
+ DeathLocations.shp: Point files listing each location where deaths were reported. It should be noted that GPS position is not being used for security precautions, and owing to the difficulty of reporting conditions on the ground, we have been using volunteers to refine location reports, but in some cases we have had less to work with. The result of this is that the grouping of locations within a small region may not always reflect the physical deaths at the neighborhood level.
+ Governates.shp: Governates (Muhafatha in Arabic), are the first level of Administration in Syria-- there are 13 provinces, and Damascus City. The province around Damascus City is known as Rif Damasq in English. We have region files, and many attributes for deaths in each.
+ Admin3.shp: This file has shapefiles for the Level 3 adminsitrative units (Nahya) in Syria (approx 270) and totals for each based on a mapping to Death Locations.
The shapefiles we used were constructed by adding attributes to those found here: http://cod.humanitarianresponse.info/country-region/syrian-arab-republic
This is an excellent site, and worth investigating if you want shapefiles transportation routes and other data.
In the absence of a formal data dictionary, here is some general guidance on the attributes found in teh DeathLocations, Governates, and Admin3 shapefiles:
Any attribute ending in AR is an arabic text string. Those ending in EN are in English. Governates have a two charachter ID string. Admin3's have a six character ID string.
A "Montika" is a Admin 2 level region. We have not found the need to generate shapefiles at this level.
Any attributes for the Governates and Admin3 shapefiles that refer to Total Population, Male Population, or Female Population are drawn from the 2004 Syrian Census. (We don't have any more recent sources.)
Deaths Per 10,000 population is the total of deaths reported (in 2011-present) for every 10,000 of 2004 population. While this measure is not precise, the aim is to get a notion of which areas have suffered more on relative basis. (Not only does this not measure any population gains in an area, but it does not address refugees.)
Deaths are also broken into male or female totals, and the Perc_Female is the percentage of female deaths to the total deaths for the region in question.
the attributes referring to "Army" do not refer to the FSA, nor do they refer to armed forces supporting the Assad Regime, but rather, defectors from the government armed forces.
Ages have only been specified in approximately 30% of all the reports-- so the average victim age should be taken with a grain of salt.
Last30, Last60, Last90-- total deaths in the last 30, 60, or 90 days.
Totals are also given by cause of death:
- Beat_Stab (beatings, stabbings and other battery)
- NotObey (Executed for Not Obeying Orders)
- Stress (Typically heart attacks in response to threatening)
- Wounds (Death from wounds, likely due to lack of access to suitable care.)
- Asphyxiate (choking, gassing, etc)
- Outage Power outages in Nurserys and Care facilities, or lack of access to care)
- Crushing (Typically by tank or other vehicles)
- Unknown: No cause of death specified-- mostly likely gunshot wounds (but not due to a sniper attack) Note that the rate of unspecified deaths has gone down a lot in the last few months
- GunandUnk: Sum of gunshot and unknown deaths. (This is most likely the best measure of Gunshot deaths, although it may run a little high)
A couple other counts:
- Shabiha: Count of deaths were the shabiha (paramilitary, militia, or thugs, depending on your preference) were involved. This is based on a text search of the comments field. This may not be the most representative measure-- some of the major massacres attributed to the Shabiha have not been reflected in the comments.
- Christian: Times when the term Christian is mentioned in the comments field
- (Note-- Use of the comment fields in our reports do not consistently report if the victim was a member of the Alawite sect that occupies so many key roles in the Assad Regime. Also, counts of Alawites have been conspicuously absent from the Syrian Census for the last few decades.
- Under 18, count of deaths where the age is specified AND under 18. (Note-- We may actually be inverting this count-- need to check)
- SameBirth measures the victims who died in the same governate (Muhafatha) where they were born.) This is working for all governates except for Dayr Az Zor, where we need to do more
- There are several attributes we are tracking for some data quality measures--- these are primarily for internal purposes, but we are including them nonetheless.
- wName-- count of victims where the name has been provided.
- wComment: Count of victims where the comment field has been filled out.
- wVideo: count of victims were a video link has been posted. (Note-- the totals may be higher-- for our reports, we've not always had the time to pull out the video links. So this number may be underpresenting.)
- wPhoto: Same for links to photographs.
- Finally, there is a series of attributes of monthly total deaths from March 2011 through the present for the location or region-- these can be used to generate "time sliders" to show different intensities over time.
Please directed any questions you my have to firstname.lastname@example.org.