Chapter 2: Geospatial Mapping and EA Database Management
Namibia
Namibia aimed to provide detailed information regarding the use of built-up structures as well as develop a Statistics Business Register and the National Dwelling Unit Frame meant to provide detailed household and land use information for each dwelling unit.
The approach to questionnaire design focused on capturing the main elements relating to the use of a building and in the case of a dwelling unit, occupants were asked the name of the head of the household, contact details, the number of rooms, and the number of usual members residing in each dwelling unit including other basic services regarding the structure they reside in, such as the source of water, electricity, and access to sanitation. The inclusion of the many stakeholder requests changed the project from a typical cartographic exercise to a statistical undertaking.
The application used to collect, transmit, store and clean mapping data was designed and developed using ArcGIS Survey123 by Esri.
The main objective of the pilot mapping exercise was to test whether the mapping tools (electronic data collection application and questionnaire) were adequate to provide the required data. This also involved testing the adequacy of logistics and administrative arrangements on the ground.
The results from the pilot mapping were used to review and improve on the mapping implementations in all operational areas. These included reviewing of tools, drawing up the final plans for the main mapping project, as well as providing the final estimation of resources required.
A Communication Plan focusing on advocacy and publicity of the mapping was developed. The most convenient method used was the distribution of flyers and displaying of posters to create awareness. Engaged community leaders through meetings and had the opportunity to elaborate on the objectives of
the census mapping. Mobilisation was done in each selected EA before commencement of data collection exercises to ensure that local people were aware of the project and what was expected of them. Courtesy visits to constituency and local councilors were also undertaken to introduce the exercise and request them to inform their constituency inhabitants about the exercise during their respective radio announcements and community meetings. In addition, FM Radio announcements, local newspapers, television, and specific talk shows among others to announce the commencement of fieldwork.
Kenya aimed to update their geography file for all counties and sub-counties with villages and EAs, as well as the number of households.
Two household listing forms were developed: rural homestead and household listing form (F-54-5-1A and F-54-5-1B) and urban structure and household listing form (F-54-5-2) used for collecting data in rural and urban areas respectively, both collected the Global Positioning System (GPS) coordinates for each household and their elevation. In addition, there was a summary form used by village elders (F-54-5-1A) while F-54-5-1B was used by the data collection teams to compile villages and EAs.
The Open Data Kit (ODK) was a free application used to develop and test the mapping data collection system and later, the ArcGIS Survey123 app, managed by Esri- Eastern Africa, was adopted after being tested and piloted. Although ODK was easy to operate, compatible with Samsung J2, and freely available in the Google Play store, Survey123 was introduced because it was faster to process the data in the ArcGIS environment being used in map production. Survey123 also provided a dashboard for monitoring teams and individual Mapping Assistants (MAs). After the completion of a village or block, the MA would edit and upload the data to the cloud server.
The technology and instruments/tools embraced were tested before the actual cartographic mapping exercise. A pilot mapping exercise was conducted to identify the most ideal tablet for use, especially one with high GPS accuracy. The Samsung Galaxy J2 phone was identified as having accurate GPS readings and tested withthe ODK application, the use of Satellite imageries and aerial photographs which were used for planning and allocation of work, plotting of boundaries and features, the adequacy of the data collection forms; and the suitability of the KNBS Server and entire ICT system in undertaking the digital cartographic mapping. The pilot was also used to ascertain the type and number of vehicles to be used as well as the resources that could be required to undertake cartographic mapping in the entire country.
The publicity and advocacy for the cartographic and mapping exercise was cascaded and done through the print media, radio, TV, social media and holding physical community mobilisation meetings for the local administration who were to hold community level sensitisation meetings. Sensitisation of the local administrative leadership was done by NSO staff sensitising the local administrative officers in each county. The sensitisation meetings were held, and the administrative officers were taken through the importance of census and the activities involved in the census process, inform how the village elders were to fill in the mapping form 1 provided to them in the rural villages. Due to lack of clear understanding of mapping concepts in Kenya, there was a variance in number of villages provided and the actual number of EAs during the sensitisation exercise.
In Kenya, the exercise started with 12 teams with each team comprising of a team leader and 7 mapping assistants. There were also coordinators in charge of several teams. The number of teams was later increased to 40 with a total of 400 temporary Mapping Assistants who were recruited and trained to speed up the mapping process.
A team was allocated between 2 and 3 vehicles depending on the number of MAs. All teams were working either in one or two counties at a given time. A team was allocated several sub locations to map the homesteads, structures, households, and points of interest. Each MA was assigned a village or a block to work in at any given time and they would move around the entire village with the help of a village elder. The MA would use the form earlier filled in by the village elder to confirm the details listed and pick coordinates for the structures, households, and features. Data was collected using mobile phones with an in-built GPS.
Mapping for the first 27 counties was carried out using ODK App while the rest were mapped using ArcGIS Survey123.
A team of GIS technicians were recruited to support the GIS team in undertaking map production.
Stage 1
-
Digitisation was carried out by first downloading the spatial data collected from the cloud server. ArcGIS software was used to overlay spatial data on satellite imagery for rural areas or aerial photographs for urban areas. The data would be symbolised by village name to determine the extent of a village. Digitisation of village boundaries was guided by already generated schema.
-
Delineation of EAs was done based on measure of size which is between 50 and 149 households making an average of 100 households. A village with more than 149 households was subdivided to create more than one EA whereas one with less than 50 households was merged with another next to it. An EA was therefore either part of a village, an entire village or several villages combined.
-
The EA polygons were cleaned by creating topologies to show where overlaps and gaps exist. Each EA was described by its administrative and political units, type (rural or urban), status (informal or formal) and universal (special or conventional). The information was added on the EA polygons attribute table to create the census geo-file.
-
A 13-digit code was generated from administrative units to uniquely identify each EA. Maps were prepared using templates generated during development of tools. Two maps were created, a sub-location map with all the EAs or several EAs in that sub-location and every individual EA map.
Stage 2 The maps were printed for editing guided by the editing forms and later printed for use during verification. Rural maps were printed on A3 size paper while urban maps were printed on A2 size paper.
-
Maps returned from verification were returned to map production for the GIS technicians to effect changes from the field. After effecting the changes, final PDF maps and geo-files were forwarded by counties to uploaded on the tablets to be used during census enumeration.
-
Hard copy maps were also printed for training and EA identification.
The EA map verification exercise was undertaken by trained verification officers and coordinators who worked with the National Government Administration Officers (NGAO) with a collective role of confirming that all the maps and geo-file for their respective administrative units were available and that they were in their correct geographic locations and had correct spellings. During verification, several NGAOs were brought together.
Additionally, the verification process identified missing features/facilities, spelling mistakes, confirmed the number of villages and boundaries, etc. The key tools were therefore the sub-location map and geo- file. The exercise started after several EA maps were produced for several counties. It was conducted in phases where in each phase several counties were covered.
Verified maps and verification forms were forwarded to the map production office to effect the changes documented before final production of maps. Major errors noted during verification included: spelling errors, wrong village/EA naming, incorrect EA numbering, disjointed/fragmented (flying) villages, omission of features, gaps where some households had not been mapped, unclear boundaries, incomplete naming of features. 32% of the EAs had one or more of the above errors.
Sierra Leone aimed to achieve a verified and categorised nationwide locality list of all administrative levels showing the name of each locality, district, chiefdom, and section with a unique code.
The cartographic mapping questionnaire was developed to capture information on, building use, locality class, classes of roads, water bodies and expected drainage type etc. The questionnaire was mainly designed to enable the head of the household to give information about the structure and number of members in the household. Other information was obtained using observation methods by the field staff.
An android-based mapping portable android device (PAD) application with a data acquisition package was built by Milsat Technologies to be used for the cartographic mapping exercise to fit all the data needs in a robust manner. The application was able to prompt the mapper when it reaches a minimum threshold of an EA to be delineated. It also ensured that the mapper is within the confines of the structure wanting to collect information as seen on the satellite imagery. The application integrated both the spatial data set and the data attributes collected on the same platform. The application has provision to work both offline and online that ensured that work is continuous even where there is no internet coverage. These innovations made it possible to achieve a sustainable national geographic frame for Censuses and Surveys.
The pilot mapping was done in two phases and the section administrative boundary shape file was used:
- The first phase was done in the capital city to: assess the cartographic mapping software EA PAD 4.0 and the means of data transmission; assess the use of ArcGIS software in processing the data from EA PAD 4.0 for EA delineation; test the entire systems and processes including logistical requirements, deployments, average time spent to complete an EA in both rural and urban areas; and identify likely problems to be faced by mappers in the field.
- In the second phase, the coverage was increased beyond the capital city with diversified prevailing conditions including hard-to-reach areas and poor internet connectivity. Three sections were randomly selected across the country.
The Communication and Publicity division within the NSO led the campaigns to sensitise and educate the public on the need to cooperate and support the mapping teams before the start and during the cartographic mapping process. The key tools used included community sensitisation meetings, radio discussions in both local and national languages, use of megaphones, providing local guards by the communities as part of the process to lead field staff for hard to reach areas and ensuring all communities are reached.
In Sierra Leone, the actual field work started with annotation of Satellite imagery, updating of verified locality list, deployment of field staff with required logistics and delineation of EAs using the satellite imagery. The data quality assurance officers and assistants were busy with the required quality checks. The national quality officer was supervising all the quality assurance officers and also used a customised application to ensure that data from the field mappers shows a true representation of the real information on the ground and in addition do ground trotting for randomly selected delineated EAs.
-
Mappers were in the field collecting data and creating temporary Enumeration Area boundaries and saving the data into the cloud. The GIS Officers and assistants who were in the respective District Offices downloaded data as sent from the field via google cloud. The EAs were digitized and delineated using ArcGIS Mapping software tool.
-
The mapping PAD remote submission process utilises a top-level security cloud system to scan and safely deliver uploaded tasks to GIS portal. GIS portal can only be opened by registered and verified GIS Officers.
-
No physical maps were printed for actual census enumeration; instead a Census Pad application was used with online resources like Open Street based on cartographic enumeration area boundary shapefiles, and the spatial data collected. These facilitated the census staff to locate their Enumeration areas without hard copy maps.
The verification process was followed up with a ground trotting exercise. To ensure that only accredited functionaries use the App, the mapping PAD periodically runs system authentication on every phone the App is installed.
Senegal
The design of the Mapping tools was a build up from versions of the previous census. They were finalised in technical workshops and benching marking at theNational Statistics Institute of Cape Verde (INECV) (for Cartography) and the National Institute of Statistics and Demography for Burkina Faso (INSD) (for enumeration).
The software used for development of the RGPH-5 Mapping tools were: ArcGIS and MOBAC
The cartography team included: mapping agents, team leaders, regional supervisors, and headquarters experts. Each of these had a functional mobile application under Android set up centrally for an integrated management of census operations. The supervisor application was used to ensure coordination and monitoring of field operations. The regional supervisor application was used to manage team composition, assign work, check completeness, and send to the server. The team leader application allowed the team leaders to assign tasks to team members and coordinate field work. The cartographer agent application was used to carry out cartographic data collection on the localities.
Botswana
SmartClient for Census, a Web-based Client-Server software solution, was used to collect, transmit, store and clean mapping data. SmartClient for Census enables different users to log into the system to execute tasks allocated to them and has GIS architecture with one central database for office and field work, which ensures easy access to data. This ensured that through unique workflows the data capturing was viewed and managed as soon as data was synchronised into the server and field work was ongoing. SmartClient for Census also has online and offline functionality, that ensures work is continuous even where there is no internet coverage.
Zambia
Zambia partnered with GRID3 to produce high-resolution gridded population estimates using a bottom-up modelling approach. This method integrated settlement data, building footprints, and population input data collected by the Government of Zambia. The resulting datasets provided population estimates at approximately 100-meter grid resolution, including breakdowns by age and sex. These estimates were used for microplanning in health interventions, such as malaria control and COVID-19 vaccination campaigns. The data also supported census planning and were made publicly available through GRID3’s Open Data Hub.
South Africa
South Africa conducted its first fully digital census in 2022, using three data collection modes: face-to-face (CAPI), telephonic (CATI), and online (CAWI). To support this, Statistics South Africa (Stats SA) updated its Geo-Frame across all settlements, collecting data on dwelling units, addresses, and gatekeeper information. This update helped fieldworkers navigate EAs and ensured accurate data collection. The digital census infrastructure included tools like the Digital Census Atlas and interactive mapping products, which allowed users to explore demographic data at various administrative levels. These innovations marked a significant shift toward integrated geospatial and statistical systems in South Africa’s census operations.