infographics

Mapping Human Activities and Resource Flows in Informal Settlements

Introduction
On the first of November 2016 I was tasked with creating a map for an informal settlement as part of the urban Modelling and Metabolism Assessment Research Team (uMAMA). The informal settlement, called eNkanini, is located on the south east periphery of Stellenbosch, Western Cape. It was formed in 2006 with just over 47 families (Van Breda, 2011), and according to data from November 2016, eNkanini has over 2800 iron corrugated households and about 200 basic infrastructure services and socio-economic activities, such as municipal toilets and taps, tuck shops, fast food shops and small farms. The purpose of this map was to collect data and analytics about land use, human activity and infrastructure types that influences the flow of resources in informal communities. This was vital for understanding the dynamics of eNkanini and building stronger knowledge forms about differing resource demands and flows in formal and informal settlements, since the predominant data collection and analysis is about formal settlementsi.

1) satellite image of eNkanini:
Pasted Graphic 1
At first glance, eNkanini has a dysfunctional and chaotic semblance like many unratified informal settlements in South Africa. This may not be the experienced reality of residents, , but could be my personal perception about the area - the repercussion of partiality and preconceived ideas I have about informal settlements. For this reason, when I first entered eNkanini, I had no utilitarian expectation of the community or about the settlement itself. However, this changed as I spent more time within the community. Without ignoring the ills within eNkanini, I started to see functional and pragmatic aspects of the settlement, rather than dysfunction based on personal biases.

Community Experiences
Data collection in informal settlements is a challenge and a laborious process. Firstly, informal settlements have narrow, winding pathways and shacks are concentrated in constricted latitudes. Secondly, informal settlements can be socially unstable due to cohesion privation with the broader society outside of their space. For this reason, you are more likely to face resistance in undertaking any data collection within the community, particularly if the community is not participating in the project. Thirdly, politics are a concern especially with projects akin to mapping. This is because the community associates maps with possible infrastructure development, whereas an academic may use maps mainly for data visualisation. This political predicament is acerbated by years of empty promises to deliver services to the community from political leaders once they have been voted in - hence I avoided wearing any shirt that had a colour which might have resembled any major political party in South Africa.
Besides creating a map, this process was valuable to me for getting a glimpse of the challenges faced by South African local municipalities. That is, providing basic services (water, electricity and shelter) to a dynamic population that is constantly moving for better living conditions. One major challenge is that there is no guaranteed revenue for providing and maintaining such services in informal contexts because they are not necessarily recognised as part of the urban landscape (Lemaire, 2015). This constant movement for better living conditions in the end produces more informal settlements, which, according to Huchzermeyer (2008), are key performance indicators of a national government’s ability to mitigate poverty and urgency to better the lives of the poor. This is because informal settlements can be used as a proxy for a community’s risk of floods, fire and other threats to health. This motivated to me the importance of providing full accounts of social and economic activities and access to basic services through spatial specific data.


Technical Considerations
Geographic Information Systems (GIS) has become crucial for meticulous decision making in modern societies at local, sectional and national level, due to the ability to spatially plan with accuracy. GIS data analysis and storage requires time and skill to produce accurate analysis and quality data visualisation through maps. I used handheld e-trex Garmin GPS for data collection. It had its own challenges due to the 95% location accuracy it has, which is about 5 metres in relative distance. Inherently, this imposed a small uncertainty on the data. Moreover, the clustering of different land-use activities within constricted spaces in informal dense areas contributes to this inaccuracy. Therefore, it became difficult to depict accurate and clear location points of every shack, infrastructure and land use activity occurring within the settlement. As a final product, we ended up with a mesh of location points overlapping on each other. Below is an example of how the distance accuracy of the GPS affected visualisation of data and made the initial state of location points unclear.

2) Map showing overlapping way points due to GPS accuracy:
Pasted Graphic
The most repetitive processes needed before visualisation was possible was cleaning the data and organising it into relational tables with associated attributes. This began by firstly examining how raw data might have deviated from reality and creating erroneous data relations. This was necessary in understanding how uncertainty arises and propagates through geo-data life cycle. Therefore, in this case I focused on the accuracy of the GPS and how it might affect data quality, accuracy and redundancy. Through this process, I was prompted to develop data quality strategies that considered the context and the type of data (geo-data, qualitative and quantitative), how it was stored, data flow from one format to another, and data-sharing within a group of people, requiring a database management system.
Data modelling and entity relation diagrams:

A) Data-base modelling is essential in ensuring databases become user efficient and serve the purpose they are built for. Data modelling begins from data collection to data cleaning and storage especially for complex data-bases. Below, it is a diagram that depicts data modelling processes towards creating a functional relational data-base for eNkanini.

B) Entity Relationship modelling is a top-down analysis technique which shows entities and the relationships that links them i.e. how each piece of data relates to the other as shown below. It begins with an entity relationship diagram that has characteristics that uniquely identifies an entity occurrence, in this case classification, attribute and location. Finally a table was created that allows quantitative analysis through Structured Query Languages (SQL).

Conclusion

Mapping eNkanini could have been done using satellite imagery, raster analysis and digitising main settlement boundaries without having to walk into the community. However, this has limited benefits with regards to data visualisation and analytics because satellite imagery does not provide societal experiences and further insights on self-attained infrastructure and services. From the conceptualisation of the research project, we understood the benefits of entering the community and how holistic spatial data has become a significant part of co-creating knowledge and understanding societal issues.
Furthermore, rich GIS data adds immense value to decision making with advanced analytics capabilities. Spatial analysis can be used to query large and complex data sets to understand behaviours, identify hotspots and predict future outcomes – making it easier for analysts to uncover actionable insights that will help shape communal sustainable growth strategies.

The maps produced through this process will be shared in future vignettes.

Notes

i See Smit S, Musango JK, Brent AC, Kovavic Z (2017). Conceptualising Slum in an urban African context. Cities. 62:107-119 http://dx.doi.org/10.1016/j.cities.2016.12.018 for a discussion of informal settlement types.