top of page
  • Patricia, Tommy, Elena, and Christopher

Power Mapping brings Rapid Reliable Energy to Rural Communities

Mapillary is teaming up with YouthMappers, researchers at Arizona State University, the OSM chapter in Sierra Leone, and other partners to map fundamental spatial features that speed up efforts to design and install mini-grids in rural areas. By understanding settlement patterns, road networks, and existing electrical grid infrastructure, the team can speed up and scale up approaches to design mini-grids for rural electrification.

One female on a motorcycle looking at the phone a Male is showing her. While another male appears to take a picture with their mobile phone on a selfie stick.

Ibrahim Yusuf Jalloh, Fatmata Kabia and Ibrahim Kalokoh (L to R) capturing street level images (UNIMAK YouthMappers, Lunsar, Northern Sierra Leone)

Sierra Leone, on the coast of West Africa, has a vast landscape connected by many roads leading to communities where most of the country’s population reside. Access to electricity across Sierra Leone is limited or unreliable, especially for rural communities. This has led to the exclusion and underdevelopment of remote communities with women and girls being the most affected. A new project seeks to address these issues.

Off-grid rural electrification through mini-grids are an optimal approach to addressing these challenges, but limited data on existing power system distribution networks in Sierra Leone hamper design processes. This collaborative project aims to map rural town distribution networks to inform mini-grid feasibility analyses. The methodology puts more towns on the map, as well as the roads running through and between them, then uses street-level imagery to survey the locations of utility poles that indicate connectivity to the power grid. This is being done with the efforts of YouthMappers chapters within the country, OpenStreetMap - Sierra Leone, the Mapillary team, and the ASU LEAPS team.

The data is being used to speed up the feasibility and design phase for installing micro-grids in scores of remote communities across the country. The methodology will support equitable access to electricity and is used as a model for scaling up electrification activities across the continent for the United Nations Sustainable Development Goals (SDGs).


Many rural villages and towns struggle with access to adequate and reliable electricity to meet their needs. The United Nations calls for more focused attention to increase electrification in sub-Saharan Africa and advance Sustainable Development Goal 7, affordable and clean energy for all, by the year 2030.

Women and girls are especially empowered by access to modern and affordable energy in remote communities. They often bear a disproportionate labor burden for their households, while reliable power can help to alleviate many of these domestic tasks. Entrepreneurial activities that improve women’s incomes can be accelerated in places where electricity access is currently lacking. The provision of adequate energy sources also reinforces efforts to prevent disease and fight pandemics – powering healthcare facilities and enabling communications. Mini-grids will support the electrification of these rural towns isolated at long distances from the national grid. Mini-grids are also adapted to provide electricity through renewable energy generation like solar photovoltaics (PV) and wind.

Distribution network data mapped in OpenStreetMap will be used to inform national and local decision-makers where and how to implement new mini-grids in rural Sierra Leone. By understanding the location and layout of current low-voltage distribution networks, more accurate estimates of mini-grid configurations can be designed.

Photo taken on a rode with several pedestrians walking or on motorcycles. Along the road are light poles and trees.

A main road in Lunsar showing utility poles and other roadside assets, by Mapillary user jalloh1113

The data will also be used for other electrification planning by the Ministry of Energy, national utilities, and other energy sector stakeholders. While regional transmission networks may be adequately charted, the local distribution networks are not typically well mapped. For this project, the team is focused specifically on mapping rural township distribution networks. These are the low-voltage power lines that connect to homes and businesses. Transmission networks, another type of grid network, transmit electricity over long distances between cities and towns. For distribution networks, mapping buildings helps improve understanding of where residents who use electricity live, and mapping roads indicates where power lines are found.

Drone image shows the top of several one story buildings with metal roofs, electric poles, and tropical trees.

Tommy Charles (Utility Poles, Makali, Northern Sierra Leone, SGA-YouthMappers)

The approach is an innovative combination of local and global volunteer open mapping with exponential technologies like augmented mapping, machine learning algorithms, and AI computer vision techniques advanced by Mapillary.

1) Augmented Feature Extraction

This step uses AI-supported digitization of roads and buildings from satellite imagery to locate populated settlements and transportation infrastructure. It is managed on the TeachOSM instance of the HOT Tasking Manager that links to the Map With AI tools of Facebook and Microsoft. The work is performed by the entire YouthMappers network, local and global. It is remotely validated by the YouthMappers Validation Team, led by George Washington University.

2) Street-level Field Data Capture

Low-voltage distribution lines are typically along roads, so the team is capturing cell phone images along mapped streets using the Mapillary mobile collection app. In six months, since March, YouthMappers from local Sierra Leone chapters and the OSM Sierra Leone team trekked across rural communities and roadways by motorcycle and bike. They supplemented some of the photography with drone support to capture utility pole patterns. Click here to see an example of a street-level field data capture.

3) Remote Crowdsourced Verification

To identify where utility poles are found inside the Mapillary imagery, the team relies on the automated process that uses AI-enhanced object detection, where utility poles are one of 42 features detected using Mapillary’s computer vision technology and semantic segmentation. To make sure not to confuse features, such as tall palm trees, with actual utility poles, this detection process requires verification. Verification is done with volunteers who are both local mappers and global members of the YouthMappers network, which includes chapters on 232 university campuses across 52 countries. You can take part as well using the Verifier Tool.

4) Definition of System Topology

Once the point data for utility poles are created, the building footprint polygons are marked, and the linear road network data traced, these features are extracted from OpenStreetMap. These features are then consolidated to model the existing distribution grid of townships. Models are then used to inform the design of off-grid mini-grids distribution networks for rural communities of Sierra Leone.

Topology data is also used by the ASU Laboratory for Energy And Power Solutions (LEAPS) to inform the development of a rapid mapping mini-grid feasibility analysis tool. GIS data is used by the LEAPS team to apply algorithms and translate geospatial data into power system data, load profiles, and resource availability. ASU algorithms perform clustering analysis, identifying the size and boundaries of the mini-grid, and initial asset placement. These geospatial data are used for XENDEE algorithms that allow power system asset selection and sizing, controls and dispatch, and final asset placement.

5) Power Engineering Design

Finally, with these inputs the LEAPS engineering team can design for power flow, distribution network design, and produce a validated model with a balanced distribution system using XENDEE, OpenDSS, and similar tools. Detailed technical specification from power system analysis is then able to improve mini-grid project cost estimates by up to 60%. This culminates the feasibility assessment with a preliminary mini-grid design. The results can then be mapped back into QGIS to visualize the project for energy sector stakeholders. This approach helps to engage communities, especially youth, in mapping their own communities and contributing directly to rural development. It leverages local knowledge and activity in ways that speed up the scoping process for micro-grid projects as much as 90% faster. The feasibility assessment tool also represents a viable way for scaling up solutions – the methods here will be applied to residential and commercial communities across Sierra Leone, and potentially inform development plans across other countries in Africa.

Flow chart representing the 8 steps in the ASU LEAPS rapid mapping mini-grid feasibility. 1. Data Collection 2. Feature Identification 3. Data Translation 4. Preliminary System Data 5.Topology Identification 6. Power Engineering 7. Financial Model Evaluation 8. Visualization for Planning

ASU LEAPS rapid mapping mini-grid feasibility analysis tool that utilizes the data production described in this blog, LEAPS, 2020.


Thank you to Drishtie Patel, Facebook; Harsh Govind, Microsoft; Mukhtar Hamzat, Nathan G. Johnson, & Shammya Saha, ASU Laboratory for Energy And Power Solutions (LEAPS); Rory Nealon, USAID GeoCenter; Stephen A K Kassigbie, Tigidanke Fofana, Grace Kainessie, and all of the YouthMappers students who contributed from Fourah Bay College and University of Makeni in particular; Jennings Anderson; Richard Hinton, YouthMappers and George Washington University; Dara Carney-Nedelman, YouthMappers; Marcela Zeballos, YouthMappers and Texas Tech University; Nuala Cowan, World Bank; Humanitarian OpenStreetMap Team; ASU Knowledge Exchange for Resilience.

This article was originally posted on Mapillary's Blog.

/ Patricia Solís (YouthMappers); Tommy Charles (OSM Sierra Leone), Elena van Hove (Arizona State University), Christopher Beddow (Mapillary)


bottom of page