Consideraciones finales

The geospatial data collected by the volunteers from the three peri-urban areas of Imhazim, Mahkama and Wardia was examined and analysed to compare the positional accuracy of the three different methods in relation to the official data collected by professional surveyors in the Al-Hillah land administration office. A customised dashboard tool developed in Matlab for calculating and visualising the RMSE and other measures was used to quantify the discrepancies between VGI and official data, as shown in Figure 6.16.

Figure 6.16 Calculation and visualisation tool for RMSE with VGI points captured by GPS compared to parcel coordinates from official mapping in the Mahkama peri-urban area of Al-

Hillah

The results of the positional accuracy tests for all peri-urban communities are shown in Table 6.34 below:

Table 6.34 Root mean square error (RMSE) for a sample of peri-urban community plot corners in the datasets compared

Peri-urban community No. of points tested RMSE smartphone and official RMSE iPad and official RMSE paper photo and official Imhazim 68 2.710m 1.482m 2.341m Mahkama 60 2.999 1.337 2.365 Wardia 180 2.992 1.308 2.067 Total 308 2.933 1.354 2.190

According to the RMSE results of the for the three communities, the use of the iPad achieved the best accuracy even though most of the local volunteers were not familiar in using it. The main reasons for this are that the volunteers received effective training and support from the researcher, and the ability to zoom in and out helped people to choose the scale they preferred and they were more careful and precise in following the instructions about demarcation. The accuracy of the geospatial data collected with pens and paper-printed aerial images scored an average accuracy of about 2 metres, which is acceptable as the scale of the photographs was

about 1:1500; however, the volunteers were a bit faster and less precise in drawing the boundaries.

The least accurate method was the use of the smartphone GPS, which achieved an average accuracy of about 3 metres. This was the best level of accuracy that could be obtained from the smartphone for the peri-urban communities. Although, the accuracy was not high, the measured positions could be used to give approximate locations for the specific plot. The following Figure 6.17-6.19 show the three collected VGI layers for the selected peri-urban areas of Imhazim, Mahkama and Wardia respectively.

6.8.4 Completeness

The total number of plots identified have been compared for each case study site, and a significant difference between the formal and volunteer quantities has been found in the peri- urban areas, with 223 plots on the official map and 285 plots observed by the volunteers. The formal data for peri-urban sites was more in line with what volunteers found, and therefore considered to be more complete because these areas exhibit less change than urban areas. This is because of the lower land values, which makes sub division less financially beneficial. In addition, the population is more settled and stable, with a lower rate of turnover. However, some sub-division does take place. The sub-divisions in peri-urban areas were different from those in urban areas because there is a noticeable lack of vertical sub-divisions: here horizontal (plan-based) sub-divisions were most common (see Figure 6.20 below). This might be due to the abundance of space in these communities and a lower density of built-up plots.

Figure 6.20 Two Moshtamals in Imhazim peri-urban district

The professionals were interested in the data about sub-division in peri-urban areas. They suggested that they may use such data to consider the possibility of changing the minimum size of plots that can be registered in peri-urban areas or they may use this data to plan utilities and roads. They also need accurate data about sub-divisions to help in long-term planning.

The nature and environment of the peri-urban communities were also different from urban areas because of the low numbers of land parcels which were occupied and subdivided compared to

6.8.5 Currency of data

The peri-urban local community volunteers were asked about their opinions of the official cadastral maps of their communities and the differences that they could find between these maps and current developments. There was common agreement between the local residents about some changes that had happened in their communities but where the cadastral maps had not been updated. Their opinions have been collated and are summarised in the following points. The maps were very old and do not represent the current development in the areas. For example, one whole block of parcels in the Al-Mahkama urban community which was considered to be a heritage site where no one was allowed to build on it. However, this block still appears on the official cadastral map (Figure 6.21). Secondly, some land use changes were not updated on the cadastral map. For examples, a small river channel in the peri-urban area of Imhazim had dried up and residents on both sides of the channel had occupied the area and changed it to become their own space (Figure 6.22).

Figure 6.21 A block of plots that needs to be removed from the official cadastral map

Figure 6.22 A river that needs to be removed from the official cadastral map A. On the map B. Currently as garage and empty land