The previous graphs offer some interesting data which may be interpreted in a number of different ways. In the case of the lower one, which gives us the percentage distribution of Caesar's traits for each book, we see can observe an interesting phenomenon, namely that sample size seems to be highly relevant to the output data, as some of the longer books such as 1, 5 and 7, tend to be more evenly distributed. Particularly in the case of the latter, which at roughly 90 chapters is nearly twice as long as the next largest book, we observe a very uniform distribution of the various traits. Also notable is the fact that book 3, in which Caesar largely directs the campaigns by proxy as he is not personally present for most of the action, clearly reflects the strategic skewing of this section of the text. Although less dramatically represented in the graphs, the situation in book 4 is also strategy-oriented, and is thus borne out in a similar way.

The line graph at the top is more difficult to interpret. We are able to get a more narrative-like sense of the events taking place within the books, but it is challenging to draw any conclusions about Caesar's methods without the benefit of visible information on the activities of the other actors. This is a conundrum we weren't able to successfully solve, namely, how to observe all the traits for all the actors simultaneously (and by chapter, to maintain the narrative effect) in a manner that is visually comprehensible. Going forward, I would like to find a way to both solve this design question and also better integrate it with the map feature, so that all of the available contextual information can more readily be synthesized by the user. Breaking ground on the project back in January, this is what I had envisioned; viewing the graph data alongside the map, I hoped that certain spatial dynamics which might otherwise be extremely difficult to notice throug normal reading might reveal themselves. Hypothetically for instance, perhaps there is a particular method to the way Caesar sets up his winter camps, making peculiar use of strategic topography that might be difficult to visualize when working with pure text. In this case the additional context of the trait percentages might direct the researcher towards these observations, or help to corroborate them.

The above hopefully proves interesting for users more interested in the multitude of other actors who make appearances in the Gallic Wars. Offering a few observations of my own, I'd like to draw the user's attention to book 2; interestingly we see that Caesar, the Belgae, and the Belgae Coalition all feature similar trait distributions. From a methodological standpoint this is encouraging, because it shows that with larger sample sizes, as is the case for these actors, the distribution remains fairly consistent. Perhaps these results also indicate that warfare was waged similarly by different peoples in the ancient world, and that the factors which lead to consistent Roman success were in fact subtler than one might expect; Caesar, extraordinary general though he was, is known to have averted defeat out of sheer good fortune on numerous occassions.

The large number of monochromatic bars indicates that many of these actors do not appear very often and so the data gleaned is unreliable at best. Nevertheless, the tribe that only appears ought to be considered collectively along with other infrequently appearing tribes, and we might observe that they are collaborating diplomatically with each other; indeed this usually tends to be the case. Likewise, spikes in tactics within certain stretches of chapters often indicate battles, and it might be helpful for the user (particularly, those from other disciplines) to see these sorts of events (again, mass tribal parleys and battles) communicated by the site.

Additionally, there are a number of other ancillary data features I wish to incorporate in the future. I am considering filtering out the low occurence actors described above, then forming trait percentages for the major players according to spans of five chapters. This is complicated, but achieves hybrid results presenting narrative effect, relevant percentages, and both "factions"; the data would almost assuredly be illuninating. Raw data aggregates in table form, of all the actors across all books, are something else to consider was in facy requested by a member of our sister project, the Pacific Islanders. It would not only be interesting, but relatively simple to implement. Finally, I wish to form stacked percentage graphs paired side-by-side with books as the X axis, and constituted from the aggregate values of the factions; i.e. one could compare at a glance the trait distributions of Romans and Gauls from each campaign season. The hope is that we might observe how their strategy, culture, and diplomacy influenced one other.