In working within the constraints presented to us, we were able to construct a series of maps and animations to visualize and better comprehend the complex effects of war on civilian populations in Baghdad, Iraq.  While this information is technically accesible, its pure state as purely quantitative data is inaccessible to the general public.  Our goal with this project, then, is to add to the growing visual dialogue of cartographies that challenge popular and often misleading representations of civilian deaths.  By making these deaths more visible, we aim to give agency to these populations that have been systematically disempowered through under-representation.  In the process, these maps will empower viewers with a counter-cartography of Baghdad as a vibrant, living urban center rather than an empty, distant target.

        Looking forward, we are interested to see how geographers and others will better be able to develop similar projects when more data is released publicly from the U.S. military as well as the Iraqi Government.  We are hopeful that the increase of open source data and software will continue to facilitate the creation of counter-cartographical maps.

limitations

        As discussed in our methodology, a lot of the limitations in this project were due to a lack of reliable data on the city of Baghdad and the war. Some of the attribute data that we did have, such as schools, we had very low confidence in it’s completeness. As well, being unable to find data on land use, our “residential road buffers” we as close as we could come to representing residential areas in the city, and we acknowledge that this an under-representation. The constrain of finding  spatialized data on population densities, ethnic distributions, and other social data collected at the catchment size required for analysing Baghdad, limited our capacity to preform extensive quantitative analyses.  As well, we have found that this to be a problem for the wider GIS community, and other quantitative researchers hoping to release data to the public autonomously about the impacts of the Operation Iraqi Freedom, in particular on high density civilian areas such as Baghdad. It is clear that this is a wider political problem, as COSIT reports have agreed that the lack of census data has led to “rancorous disputes over the size of Iraq's socioeconomic and demographic structure,”  and is extremely valuable to a number different governmental and non-governmental bodies.

        The other problematic nature of our data is our limited confidence in the accuracy and completeness of the Wikileaks data set. Some of the drawbacks to the Wikileaks data include:

• Under-reporting of deaths. We acknowledge that the data set is not "every death" or even close to it. The data represents every death recorded by Multi-National Forces Iraq. There is under-reporting in these data for a number of reasons including: (1) there was no Coalition or Iraqi unit around to record the death; (2) the Coalition and Iraqi units in the area were engaged in such high levels of combat that did not have time to track down every casualty on all sides; or (3) if the outcome of the incident was ambiguous.

• Changes in standards of reporting deaths. The reporting standards for Significant Actions (SIGACT) changed dramatically over time and the reporting procedure varied across units. As well, there have been considerable changes made in how reports from Iraqi units entered the data. In general, the data are likely to be more complete later in the war when reporting standards were more uniform and the integration of information from Iraqi units was better. We noticed this specifically in how time was recorded, as the formats changed considerably, making a comprehensive and accurate chronological representation more difficult.

• Limited scope of the data set. Only reviewing incidents that resulted in casualties may result in a heavily skewed view of the war. Whether a violent incident causes casualties, especially civilian ones, has a large random component and therefor our maps do not provide near enough to a complete picture from this reduced dataset. Furthermore, other critical incidents, such as bombings of infrastructure (ex. Power stations, water plants, etc.) can have equally fatal results, though they are not considered “reportable” incidents in this data set.

• In addition, the logs are recorded immediately after the events take place, by the soldiers involved, so there is a large possibility of error. It is inevitable that perceptions change under the stress of combat. Also, unlike scientists or journalists, soldiers are there to be soldiers - not record events. Most likely they regard it as "onerous after a disturbing and exhausting mission." (Spiegel International)