reas 
  
he test 
| using 
se and 
nelp of 
ie. data 
al and 
nt data 
id axis 
ind the 
ents a 
Jurban, 
4a. In 
purple- 
1at the 
can be 
ie total 
ditions 
a. The 
07 and 
> Town 
during 
ncrease 
h 4-6). 
1esburg 
vity up 
)10 the 
t be an 
nce the 
'eneral. 
»r 2009 
nberley 
period. 
d early 
ions in 
olumes 
] areas 
  
Figure 1 Additions to point data for a) commercial and b) 
residential category 
The volume of data is an important factor to consider, as an area 
may have regular updates, but if the overall data volume 
remains low then the data may not be considered as of high 
value to National Mapping Agencies. 
in 
   
I. i x | wilh A A. E Hon Sn 
i 
I = dd 1 
a) b) 
Figure 5 Comparison of additions to different thematic classes 
for a) commercial and b) commercial versus residential 
The geometries in commercial and residential categories were 
further investigated with respect to their semantics. Figure 5 
shows additions with respect to the semantics. Figure 5 shows 
additions in several other categories. The vertical axis in Figure 
5a shows the count of additions. The test areas are represented 
by the same bar colours as Figure 4. The horizontal axis shows 
the amenities, ranged from Banking (1), Health (2), Education 
(3), Religious (4), Leisure (5), Safety (6) and Postal (7). 
Figure 5a clearly shows that contributions for commercial areas 
to the class Leisure were the highest, second was Banking and 
third Religious. All the test areas had the most additions in 
Leisure with the exception of Korsten that had the same 
addition count for Education and Leisure. The second highest 
count of additions varied per area. Cape Town and 
Johannesburg's second highest count was Banking. Durban's 
second highest count was in Baking and Health. 
Figure 5b shows the count of difference in additions between 
commercial (red) and residential (green) areas (vertical axis). 
The horizontal axis shows the amenities, ranged from Banking 
(1), Health (2), Education (3), Religious (4), Leisure (5), Safety 
(6) and Postal (7). The highest count per category for 
commercial areas was: Leisure, Banking and Religious. In 
residential areas it was: Religious and Leisure. Thereafter 
Banking, Health and Education categories had the same count. 
Deletions 
Deletions for all areas were generally low with a few 
exceptions. This may be a good sign, because it could mean that 
as data has been added to the OSM database over the years, the 
majority of contributions have been deemed as being correct. 
This cannot be said for certain as OSM does not have an official 
quality control process. Mistakes are identified by the users 
only. 
Modifications 
The amount of modifications is slightly higher than the 
deletions, but generally the total modifications contributed are 
low for all test areas when compared to the data that has not 
changed between epochs. This can be interpreted as meaning 
the features were correctly added the first time. It can also be 
interpreted as meaning that the features didn't change very 
much between epochs which for most features will be the case. 
4.2 Lines 
Additions 
The same tests arcas were used to investigate the change in line 
geometry. Figure 6 illustrates the results of the additions to line 
geometry. The vertical axis provides the length (in kilometres) 
of lines added or deleted between epochs and the horizontal axis 
shows the epochs. Many additions were made in the early days. 
This could mean that line data was pulled from various sources 
(including existing data sets) to create the initial base. 
Contributions for Cape Town was high in 2007 (epoch 1-2), but 
decreased thereafter and remained low until early 2010 (epoch 
7) However, between late 2010 and early 2011 (epoch 7-8) 
additions increased significantly. This is significant as epoch 7 
is before the 2010 FIFA Soccer World Cup and 8 is 
immediately after that. In figure 6a additions in Durban had a 
slow start, but suddenly increased substantially in 2008 (epoch 
3-4) and then another big increase between late 2009 and early 
2010 (epoch 6-7). Johannesburg also had a significant increase 
like Durban between late 2007 and early 2008 (epoch 2-3) and 
then another lower increase between early 2009 and late 2010 
(epoch 5-7). Korsten had the lowest volume of increase in 
additions for the commercial category with two less significant 
increases between late 2008 and early 2009 (epoch 4-5) and 
between late 2010 and early 2011 (epoch 8-9). 
In Figure 6b additions to residential areas were gradual with a 
rapid increase for Middleburg in 2010 (epoch 7-8) and also for 
Universitas between late 2010 and early 2011 (epoch 8-9). 
Both Brackenfell and Kimberley had very low additions in 
general, with a minor increase for both areas between late 2008 
and early 2009 (epoch 4-5). 
  
a) b) 
Figure 6 Additions to line data for a) commercial and b) 
residential category 
e 
    
a) b) 
Figure 7 Deletion to line data for a) commercial and b) 
residential category 
521