1. Introduction

Research Objectives

2. Materials and Methods

Figure 1. QField Screen capture. A health facility is identified and inserted in data collection in Mumias East, Kenya, using QField and Bing basemap imagery.

Figure 1. QField Screen capture. A health facility is identified and inserted in data collection in Mumias East, Kenya, using QField and Bing basemap imagery.

Table 1. Data layers contents and data types.

Figure 2. A street picture next to its aerial imagery facilitates identification of paths where they are not directly visible from aerial imagery.

Figure 2. A street picture next to its aerial imagery facilitates identification of paths where they are not directly visible from aerial imagery.

3. Results

Data-Scarcity in Rural Context (Objective 2)

Despite the presence of a few collective and individual mobility solutions in the Global South, by motorcycles and bicycle, daily mobility is generally conducted by foot. This research allowed the collection of networks of paths allowing a realistic GIS mobility analysis of rural mobility with a finer granularity of data than existing official, or private routing data providers, such as ESRI’s ArcGIS Online Network Analysis Services or the OSM current state of the map.

At the time of this study, ESRI existing mobility network as provided as network service is broadly similar to the existing OSM data. The data scarcity of the existing GIS rural data is best illustrated in comparison to the additions made by the research team, whose methodology will be explained below. A comparative observation of what areas can be reached using ESRI ArcGIS Pro services and with the mobility network data that was collected by the research team can be observed in Figure 3. The Service Area comparison illustrates the areas covered at certain network distances of facilities (in this case schools). Conducted on two maps with lower and higher network density, the comparison illustrates how the non-inclusion of paths presents many areas uncovered, because they are located at certain distances from mains roads. Rural mobility is however mainly pedestrian and a realistic picture of local mobility must reflect the uses of rural paths and shortcuts.

Figure 3. Comparison of ESRI routing services and Field data to study rural mobility.

Figure 3. Comparison of ESRI routing services and Field data to study rural mobility.

The data collection allowed the inclusion of numerous roads and points of interest, following OSM existing mapping or community-specific items. The mapping of streets roads and paths is the largest quantity of items. Prior to the visit in Kakamega, the existing OSM database contained a number of main roads and junction, but not the rural roads and small paths that are used daily by community members. In Kakamega, for example, the number of roads collected increased dramatically the total number of roads in the Mumias East region. As the attention of data collection came closer to the daily activities of rural dwellers, the number of segments increased and their length shortened, reflecting a set of activities closer to home and pedestrian in nature.

The inclusion of the walking mode of transport in analysis of rural mobility requires the creation of thousands of line segments of roads and paths for areas of interests spread over around ten kilometres by five, as was the case in Kakamega and Birendranangar. The collection of this data allows for the calculation of daily mobility mapping based on a dense network of path. The Figure 4 below illustrates the number of new segments created in the map of Kakamega. The numbers of new segments in the new GIS data represents over 300 kilometres, which is roughly equivalent to the pre-existing data retrieved from OSM data. It is notable that the number of new segments is three times higher than the original 2528 in OSM. In turn the average length of OSM segments are much larger at an averages of 126 meters and the new segments at an average of 41 meters. This reflects how the crowdmapping of OSM focussed on the main roads, while the new data collection focussed on collecting a dense net of data within a small territory.

Figure 4. Road segments collected in Mumias East over the existing map of the area.

Figure 4. Road segments collected in Mumias East over the existing map of the area.

3.4. Measuring Access to Amenities with Collected Data

The data collection allowed the mapping of access to amenities from the different points of the wards/locations. Using diverse sources, the service areas allow to locate all areas within certain distance of a specific amenity. The example of primary schools, illustrates all buildings at every 500 meters interval distance and illustrate that a large number of children need to do significant daily walks to access education. Note that here buildings do not equate with population, as many Microsoft buildings included are agricultural facilities. It serves however as a proxy to illustrate that the proportion of children needing to do long walks to school is significant.

Figure 5. Service area map of Lusheya.

Figure 5. Service area map of Lusheya.

A similar exercise conducted in Birendranagar rural ward 13 illustrates a similar problematic. The lower populations and building density still illustrates the needs to walk for significant portions of children in the area.

Figure 6. Birendranagar Ward 13 Service area.

Figure 6. Birendranagar Ward 13 Service area.

Different types of destinations are linked with diverse needs. Access to water requires the proximity of water collection points in a closer vicinity as carrying water and the frequency of the need in a household is high, with several voyages each day. The collection exercise allowed to collect 182 water points during the exercise. A map of the water points in relation to dwellings illustrates how most dwellings are located in close vicinity of collective water points.

The density of points located outside a 500 meters service area from any community well, illustrates however the large number of dwellers that need to travel more than 500 meters (going and back) to access water at all season. As community members explained, individual water points present in most houses do not give access to water all year round, when the dry season does not allow individual access. This results in a number of travels carrying water being an important time and effort made mostly by women for basic living needs.

Figure 7. Lusheya water access service area map.

Figure 7. Lusheya water access service area map.

Validation of the Data with Communities

Data Validation by Making Maps

This research project is closely linked to the development of a GIS platform application that includes the creation of maps by enumerators and community members together. This approach of data validation stems for the tradition of F-VGI, or Facilitated VGI, where participants are provided with a simplified access to GIS software and data, to provide geographic information [18,19].

This approach as conducted in Birendranagar, Nepal, allows the participant to access their data in a common exercise and produce maps from the data inputs. This closer look at the data results in maps of sufficient quality with the use of symbology discussed then in common with the community members.

Maps produced by community members belong to a GIS-amateur genre. With a simple and easy to use tool developed during the course of the project, the online mapping tool Usafiri (https://usafiri.io/), community member were invited to build themselves the maps with the support of the research team. This method had the compared advantage with the paper map exercise to allow community members to directly verify and correct the data that was corrected in the previous weeks.

A projection of the maps in a public hall that included the community participants and the elected official of the wards and municipality allowed to discuss the data collected while directly correcting the data.

Figure 8. Community map highlighting mobility barriers.

Figure 8. Community map highlighting mobility barriers.

Figure 9. Community map focussed on social infrastructure.

Figure 9. Community map focussed on social infrastructure.

Dynamic GIS Software Combines with Paper Maps

The nature of the data collection tool used for the exercise allowed the continuous control of the technical quality of the data. As QField permitted the visualisation of the collected data on the device, and later when coordinated allowed several checks of on the computer, the tool itself allowed an important assurance that no collected point or line would be misplaced. The most important quality check of this exercise is not however related to the control of point and line position, but to ensure that the main elements defined at the beginning of the collect would correspond to community knowledge.

A form of quality control was continuous, as community members ad enumerators constantly exchanged views and community members were shown the data on the device and actively commented on the situation and positions of elements. The validation needed however a specific session where the community members had access both to a wide overview of the areas and seeing all the details on the map. To allow such exercise to take place, the research team designed and printed two large paper maps in A0 format representing both the Lusheya and Khauga areas.

The validation session consisted in two main parts. A first session aimed at collected collectively the impressions of the participants on the overall exercise and whether they considered the focus of the research to correspond to their mobility priorities and issues. A second part focussed on the detail of the maps by hearing the movements of participants, and correcting some errors of positions and names of features on the map.

The discussion around the maps consisted in the participants listing all the main points of interest listed on the maps as well as adding any element that would not be present, or correct names that were wrongly spelled on the maps. This validation serves in both correcting possible errors but more broadly ensure that community participants recognize the places and validate the overall results.

Another exercise was conducted to have community members describe their own mobility patterns on the map. This part of the exercise allowed going back to the daily mobility patterns, where community participants describe their daily life, only this time it was visually supported by the map. This not only allowed to see patterns of movement of proximity, but also to evaluate what are the minimum daily distances made by community members.

Community members indicated the approximate location of their homes on the map and the main sites of their daily mobility. As explained by these community members, they made all of their daily movements walking. Public transport solutions, including three-wheelers (boda boda in Kenya or auto in Nepal) and minibuses (matatu) was done to reach areas of further distance and less frequent destinations, such as health centres or further to the central administration (Shianda town in Kenya, Birendranagar in Nepal).

This discussion of the daily activities of participants allows observing what are the activity spaces of rural dwellers. Activity spaces are defined as the convex hull of the daily mobility [20,21]. The advantage of this method is its provision of daily activity spaces as a simple area that simply requires the location points of where most activity happen.

In the present case study, they also have the advantage illustrating the area coverage of the participants familiar space over the whole study area. The map of the daily mobility hull illustrates that most activities are restricted to a relatively small area. It reflected interestingly in the validation session as several participants noted that the exercise helped them see their village in a way they had not thought earlier.

The spatial spread of geographic elements in the villages illustrates the character of subsistence economy of the areas covered during the community cartography. The need to have access to close amenities in daily life by local dwellers results in the villages being covered by dense net of small shops and industry, such as grain mills, worship places and water access, among others.

Community Members Feedbacks

A final focus group was organised with community participants to understand their opinions on the overall exercise. It aimed to understand their opinions of the exercise in a broader scope than the maps and data collected. It also aimed to let participants voice their critiques of the exercise, positive and negatives to enhance future exercises in

Perhaps the most interesting result of the exercise, is that community members pointed to how the data collection exercise pushed them to reflect on their villages and social and spatial dynamics. Some community members pointed how this exercise was the occasion to discuss in depth with the local authorities and helped them get a better sense of the mobility dynamics of their village. “I learnt to interact with people from different positions”, explained one participant.

As parts of the exercises were done walking with village elders, the participants enjoyed the deep knowledge of the elders in explaining village social and economic dynamics. Village Elders are elected semi-official acting as interface between the government and citizens, they are generally older people with deep links to their community [22]. Participants thus learnt about locations and local points of interest in the villages as well as walking and learning form places located deeper into the villages than they usually do, apart of their own immediate vicinity.

One participant noted in particular that he only now realised the importance of the regular location of boreholes and river access points. As many houses have direct access to shallow wells, these are generally useable only during the rainy seasons, while in the dry season the shallow wells do not provide enough water all year round [23,24]. The participant noted that the wide presence of shallow wells contrasted with the more spaced and less frequent permanent water points.

Participants also commented that the visits combined with data collection made them realise that clustering of some of the local amenities. “Most churches are located where most other amenities are. But it also made me realise that a number of villages don’t have their own schools and children must walk further”, noted a participant. Coupled with the observation of the map, this exercise of validation of the data and the exercise convincingly illustrates that participatory exercise have a useful effect of supporting local community dwellers to orient their observations of their own living space.

Important critiques of the exercise by community members concerned mainly the practical constraints of the exercise. The transect method requiring walking long distances over the day implied that some of the daily exercises lasted from 9am to 3pm. The community participants noted that they would have favoured expanding the exercise in each of the location to 4 days instead of 3 to allow for shorter days and distances covered per day. “You should take more time to get good information”, as one participant noted at the end of the session.

The impression on the exercise however was not unanimous among the participants. As some had suggested to widen the area of study and others on the contrary suggested surveying smaller parts and focussing on a reduced number of villages and sites. It is not surprising that different opinions emerge from the debriefing of a collection exercise. It provides key insights for researchers to understand that different methods may seem the most appropriate and that decision should take into account and weight in a number of factors.

Suggestion of broadening the scope of the survey also came in the discussion. As the Chief of one of the location pointed: “These maps can support the administration, by showing the locations of MPESA (mobile money) points, or the posho mills (small maize mills). However, when “the question of road access is a big issue that relates to the land, and to the sensitive issue of land ownership”. The feedbacks form the community members highlight numerous questions on how to conduct these type of collection exercises. They also mark eloquently how limited the granularity of official data collection is in rural Kenya.

Results of the first exercise in Kenya provided feedback to enhance and systematise the exercise in Nepal. By this time, the method integrated a GIS Web platform that allowed community participants supported by enumerators to discuss the results in smaller groups and directly navigate interactively the data and produce maps selecting their elements of interest.

Large paper maps and screening produced in advance allowed the representatives of the community and local authorities to comment on the collected data. A survey of appreciation conducted using permitted to assess the impressions of the participants on the exercise. Some participants found the exercise too long and demanding with long walks under high temperature. Validation exercise is also the occasion to assess and discuss the conditions of the community mapping. Among the remarks from participants, comments regard the level of comfort provided during the collection exercise, including for instance discussing the temperature during hall exercises or the quality of the food that was provided. Most participants noted, however, that the exercise was an occasion to observe and understand closely their environments and discussing the conditions of various amenities in their villages with local authorities.

Validation session allowed the addition of elements used in the first fieldwork with other methods. Refining the first exercise allowed to include a broader group of stakeholders in both rural and urban mobility.

The large maps that were shown to communities were commented and annotated to illustrate origins and destinations in the community daily life. In later confirmation exercise, the homes (origins) and main destinations were located on the network and calculated along the network with an Origin-Destination Matrix to see distances involved in daily mobility.

A female participant from Lusheya identified some key destinations as the Market, the Water point and her local church. Although there are numerous churches on the whole territory, the different faiths of the congregants do not necessarily to the closest to their houses. The calculated distances between origin (home) and destinations are shown in Table 4.

The distances involved in the movement of this woman illustrates well different ranges of distances depending on activities. The closest water point is at a walking distance of 310 meters and is clearly the closest distances, whereas shops and especially church can be located at longer distances. Gendered roles in the community, where water collection is almost exclusively assigned to women explains why female respondents present this as regular activity whereas male respondents do not include it in their activities.

A male member of the community, working in a local office indicated his regular movements. The distance to the church is also the longest of its regular movements, justified by a lower frequency of movement.

Figure 10. Detail of an annotated community map of Lusheya.

Figure 10. Detail of an annotated community map of Lusheya.

The table of movement samples show a close correlation of distances as estimated by the communities in the first exercise described above, where community members indicated their daily, weekly, and less regular mobility in their community. Both examples of mobility shown above illustrate relatively long walks for religious service, despite a strong density of places of worship in the area. This is due to a great variety of churches, including Catholic, Anglican, various protestant faith and a small number of mosques. As a result, on a weekly basis, church-going can involve relatively long distances.

Overall the examples calculated with network analysis are quite congruent with the estimates made by the community at the begging of the exercise as shown in table. The use of paper maps is however complex and time-consuming to re-import into a GIS. This process, as Seeger notes in his commentary: “The value of the local knowledge to the planning process [is] clear, but so too [are] the limitations of the process in which the information was collected.”[3]

4. Discussion and Conclusions

References

1. Munyaka, J.-C.B.; Chenal, J.; de Roulet, P.T.H.; Mandal, A.K.; Pudasaini, U.; Otieno, N.O. Multi-Level Participatory GIS Framework to Assess Mobility Needs and Transport Barriers in Rural Areas: A Case Study of Rural Mumias East, a Sub-County of Kakamega, Kenya. Sustainability 2023, 15, 9344. [CrossRef]
2. Bryceson, D.F.; Bradbury, A.; Bradbury, T. Roads to Poverty Reduction? Exploring Rural Roads’ Impact on Mobility in Africa and Asia. Development Policy Review 2008, 26, 459–482. [CrossRef]
3. Seeger, C.J. The Role of Facilitated Volunteered Geographic Information in the Landscape Planning and Site Design Process. GeoJournal 2008, 72, 199–213. [CrossRef]
4. Brottem, L.V.; Coulibaly, B. The Geography of the Bottom Billion: Rural Isolation and Basic Service Access in the Republic of Mali. Eur J Dev Res 2019, 31, 1147–1170. [CrossRef]
5. Berg, C.N.; Blankespoor, B.; Selod, H. Roads and Rural Development in Sub-Saharan Africa. The Journal of Development Studies 2018, 54, 856–874. [Google Scholar] [CrossRef]
6. Beuran, M.; Gachassin, M.; Raballand, G. Are There Myths on Road Impact and Transport in Sub-Saharan Africa? Development Policy Review 2015, 33, 673–700. [Google Scholar] [CrossRef]
7. Bryceson, D.F.; Mbara, T.C.; Maunder, D. Livelihoods, Daily Mobility and Poverty in Sub-Saharan Africa. Transport Reviews 2003, 23, 177–196. [Google Scholar] [CrossRef]
8. Madsen, L.M.; Adriansen, H.K. Understanding the Use of Rural Space: The Need for Multi-Methods. Journal of Rural Studies 2004, 20, 485–497. [Google Scholar] [CrossRef]
9. Sexton, A.E.; Hayes-Conroy, A.; Sweet, E.L.; Miele, M.; Ash, J. Better than Text? Critical Reflections on the Practices of Visceral Methodologies in Human Geography. Geoforum 2017, 82, 200–201. [Google Scholar] [CrossRef]
10. Hirt, I. Mapping Dreams/Dreaming Maps: Bridging Indigenous and Western Geographical Knowledge. Cartographica 2012, 47, 105–120. [Google Scholar] [CrossRef]
11. Hirt, I. Indigenous Mapping: Reclaiming Territories, Decolonizing Knowledge. In The Politics of Mapping; John Wiley & Sons, Ltd., 2022; pp. 155–185 ISBN 978-1-119-98675-1.
12. Santos, A.P.; Pessoa Colombo, V.; Heider, K.; Rodriguez-Lopez, J.M. Comparing Volunteered Data Acquisition Methods on Informal Settlements in Mexico City and São Paulo: A Citizen Participation Ladder for VGI. In Socio-Environmental Research in Latin America: Interdisciplinary Approaches Using GIS and Remote Sensing Frameworks; López, S., Ed.; The Latin American Studies Book Series; Springer International Publishing: Cham, 2023; pp. 255–280. ISBN 978-3-031-22680-9. [Google Scholar]
13. Choplin, A. La mondialisation des pauvres: loin de Wall Street et de Davos; La république des idées; Seuil: Paris, 2018; ISBN 978-2-02-136652-5. [Google Scholar]
14. Citizen Mapping: Using Common Geospatial Technologies to Map Informal Settlements in São Paulo, Brazil.; Pessoa Colombo, V., Ed.; 2018.
15. Johns, F. #Help: Digital Humanitarianism and the Remaking of International Order; Oxford University Press, 2023; ISBN 978-0-19-764887-2.
16. Brown, G. A Review of Sampling Effects and Response Bias in Internet Participatory Mapping (PPGIS/PGIS/VGI). Transactions in GIS 2017, 21, 39–56. [Google Scholar] [CrossRef]
17. Verplanke, J.; McCall, M.K.; Uberhuaga, C.; Rambaldi, G.; Haklay, M. A Shared Perspective for PGIS and VGI. The Cartographic Journal 2016, 53, 308–317. [Google Scholar] [CrossRef]
18. Seeger, C.J. Mapping and Visualizing the Built Environment and User Interaction Through Facilitated-Volunteered Geographic Information (F-VGI). In The Virtual and the Real in Planning and Urban Design; Routledge, 2017 ISBN 978-1-315-27024-1.
19. Kalvelage, K.; Dorneich, M.C.; Seeger, C.J.; Welk, G.J.; Gilbert, S.; Moon, J.; Jafir, I.; Brown, P. Assessing the Validity of Facilitated-Volunteered Geographic Information: Comparisons of Expert and Novice Ratings. GeoJournal 2018, 83, 477–488. [Google Scholar] [CrossRef]
20. Sherman, J.E.; Spencer, J.; Preisser, J.S.; Gesler, W.M.; Arcury, T.A. A Suite of Methods for Representing Activity Space in a Healthcare Accessibility Study. International Journal of Health Geographics 2005, 4, 24. [Google Scholar] [CrossRef] [PubMed]
21. Stewart, T.; Duncan, S.; Chaix, B.; Kestens, Y.; Schipperijn, J.; Schofield, G. A Novel Assessment of Adolescent Mobility: A Pilot Study. International Journal of Behavioral Nutrition and Physical Activity 2015, 12, 18. [Google Scholar] [CrossRef] [PubMed]
22. Mutua, J.N.; Kiruhi, T.M. Volunteer Public Leaders’ Values-Driven Leadership: The Case of Village Elders in Kenya. Heliyon 2021, 7, e06411. [Google Scholar] [CrossRef] [PubMed]
23. Rutten, M.M.E.M. Shallow Wells: A Sustainable and Inexpensive Alternative to Boreholes in Kenya. ASC Working Paper Series 2005. [Google Scholar]
24. Turman-Bryant, N.; Nagel, C.; Stover, L.; Muragijimana, C.; Thomas, E.A. Improved Drought Resilience Through Continuous Water Service Monitoring and Specialized Institutions—A Longitudinal Analysis of Water Service Delivery Across Motorized Boreholes in Northern Kenya. Sustainability 2019, 11, 3046. [Google Scholar] [CrossRef]