After the field survey carried out in July 2017 part of the team was trained on the identification and classification of the ceramic material picked up and bagged at each collection/sample spot in the field. During the cataloguing of the pottery the then created database of surface material has been converted into an Excel spreadsheet with the different classes and types of ceramics associated with the spot where they were collected in the field. Having already imported the GPS points, of all the collection spots, into GIS, we discussed the best way of linking the excel table to the survey points in QGIS. The information that we want to display across the surveyed area regard the different types and sub-types of material and their chronologies (where determined). As for each collection spot (represented by a GPS point) there are more than one potsherd, the best way to link the GPS points and the table of pottery types is through a relation one-to-many (parent-child) so that every potsherd catalogued per collection spot is represented and searchable. In this way, the two tables are linked and each type or class of pottery is searchable and displayable on the map across the surveyed area.
With the data stored in this way it is now possible to search for a specific category, class or type of material (from the table containing the pottery classification) and then display the quantities of the specific queried class of objects across the surveyed area (Fig. 1).
Subsequently, we introduced a new way of representing and storing the chronologies of potsherds (where determined) into the attribute table of the GPS survey points. We used an established method where time blocks of 50 years are used to store the information regarding the chronologies, as shown in Fig. 3. For each collection spot, some potsherds have been dated with different chronological resolutions (from 2 or more centuries to 25-year accuracy). Therefore, in order to overcome the discrepancy between diverse chronological accuracies, the time blocks method has been discussed with and proposed to the trainees as a way to record the chronologies of the field survey data.
For each potsherd that has been dated we counted 1 quantity in each half-century (50 yrs) field – so that if a sherd was dated 400-700 AD, a unit is counted for each half-century from 400-450 to 650-700. In this way, it is easier to display how much each half-century is represented in each collection spot. Successively, a selection of the half-centuries that someone wants to aggregate into a single layer is possible by simply selecting all the half-century with value greater than 0 and then export the selection as a separate shapefile. In Fig. 2 is shown how it is possible then to display the collection points where the field survey yielded material datable to the 6th and 7th century AD.
Day 6 was dedicated to the creation of map layouts from GIS, which allows the trainees to produce outputs that they can use for communicate and disseminate their work. This entailed the use of north arrow, scale bar and legend. As some of the trainees were already experienced with the use of the Print Composer tool of QGIS, they took the lead and taught the other participants how to use this tool, under our supervision. One of the main aims of the overarching Training in Action project is to form a number of people that can then train others after the end of the project, and therefore it is important that trainees develop teaching skills as well, starting from what they are more familiar with (Fig. 3).
The tools and skills presented and discussed in the first 3 days of GIS advanced training were then recalled during the second week when advanced photogrammetry and condition assessment have been covered.