Identifying Trade Trends in Greek Pottery

Access the GitHub repository for the project here

Note: This project was developed as a term paper study for the graduate level course CSCI-6443-Data Mining under Prof.Bellaachia. The following is the abstract from the final term paper submission. Access the full text version of the article here. View a brief presentation here.

In this case study, the KModes clustering algorithm has been applied to the data across different timelines helped identify trend groups across different geographical locations and the shift and spread in pottery techniques. The case study works with data obtained from the Beazley Archive Pottery Database (BAPD) which is the world’s largest database that contains information on ancient Greek painted Potter. The BAPD is currently being maintained by the Classical Art Research Centre at Oxford University. The dataset is composed of information on ancient vases most of which are estimated to have been created during the period from 6th to 4th century BC. For the present case study, a subset of this data containing information about the vases was worked with.

From the results of the clustering exercise, it was seen that in the period from -550 to -500, there are two main groups of Greek vases identified. The first is Pyxis shaped vases from Italy with no inscriptions and black figure painting and the second is Astragalos shaped vases from Greece with no inscriptions and black painted figures suggesting similar styles but with distinct characteristic elements. This is indicative of cultural and goods exchange between the two major identified communities – Archaic period Greece and Italy. This observation is also backed by the existence of external evidence suggesting that while these were distinct communities, significant trading was underway between them.

The clustering results of the period from -500 to -450 shows a sudden jump in the identified pottery clusters. It indicates now that in Italy, Astragalos shaped pottery with red painted figures and no inscriptions was prevalent - a shape and design characteristic of Greece in the previous year frame. Within Greece itself emergence of new shapes and designs can be seen indicating diversification of pottery production techniques. The next time period from -475 to -425 is indicative of a decline in Krater shaped pottery which is an ewer like vessel used for mixing vine with water. This could be indicative of a shift in social practices and preferences. Lastly, from -400 to -300 BCE, the emergence of Red-figure Astragalos from Spain and Redfigure Pyxis from Greece shows the further diversification in pottery styles possibly indicating the dynamic nature of artistic expression and cultural diffusion across different regions within this time period.

Thus, in conclusion it can be said that, the clustering analysis of ancient Greek pottery data reveals distinct patterns in evolution, indicating cultural exchanges between Greece and Italy, shifts in artistic preferences, and diversification of production techniques over various historical periods. The clustering results offer valuable insights into the dynamic nature of ancient societies and their interconnected artistic traditions reinforcing the utility of data mining techniques in the field of archaeology for pattern detection and hypothesis support.