Increasing Clean Energy Access in Africa Through Predictive Modeling

Increasing Clean Energy Access in Africa Through Predictive Modeling

NeedEnergy is an energy-tech startup to provide sustainable and clean energy solutions. In this two-month Omdena Challenge50 technology changemakers collaborated to develop predictive models for designing solar rooftop installations and gas pay-as-you-go reticulation services.


The problem

Sub-Saharan Africa has over 600 million people without access to electricity and electricity demand grows at an annual growth rate of 11%, the highest rate of any region worldwide. The number grows to over 700 million if clean cooking energy sources are considered as most people still rely on firewood and charcoal for their day-to-day cooking. These are just a few of the many additional challenges: 

  • The Grid is getting old and results in increased maintenance and operation cost.
  • Cost for unplanned maintenance and unforeseen faults is a pain for utilities and results in loss of revenue.
  • The Grid has not fully migrated to the edge or cloud to benefit from industry 4.0.
  • Data is in abundance but most of it is not utilized, a potential to start solving the above mentioned.


Electricity demand for commercial spaces will grow to 390 TWh by 2040 and 70% of this demand will be covered by renewable solar PV energy. This sector will experience one of the biggest energy transitions and an opportunity for a more m modern architecture for the grid of the future.


The project outcomes

NeedEnergy intends to use predictive analytics for designing solar solutions or clean energy solutions for clients based on their projected energy usage/profile. This will help to increase energy adoption where it is most needed.

You will help to accomplish this by leveraging NeedEnergy`s network of smart energy monitors for both electricity and gas. This will help with decision-making for Commercial and Industrial (C&I) clients who are transitioning to renewable energy. The analytics insights will also be used for energy suppliers. For example, gas suppliers can better plan deliveries and inventory based on the data.

In this project, you will also build predictive models to detect anomalies in the operation of the installed solar asset. An integration with IBM Deep Thunder will be ideal so that weather influences on the installation can be put into perspective when designing or operating the solar installation. 


The data

For the project, the data is classified into two main buckets, which we will use to varying degrees depending on how the project unfolds:

  • Historical Data (realized data) – This information contains the highest signal-to-noise ratio and high relevance but is expensive to collect both financially and timely.
    • User data obtained from smart meters onsight.
    • Demographic information obtained from public entities like the local utility and Regional Power Trading Data (research paper to be shared)
  • Forward-Looking Information – This information is used to provide a broader context for prediction purposes and improve accuracy when dealing with new/unseen situations. It takes into account things that may not appear in historical data sets.
    • Weather information
    • News
    • Trading Prices


The Omdena team built internal databases to store this information (relational and time series) and also develop an API to allow for easy access in production and for research purposes.


Streamlit interactive dashboard showing short-term and long-term energy demand - Source: Omdena

Streamlit interactive dashboard showing short-term and long-term energy demand – Source: Omdena


You can view and explore the dashboard using this link. To read more about how the data was collected till how that dashboard was built, please check the articles below.


Need Energy about the AI Challenge results



Increasing Solar Energy Adoption Through Roof Detection

Increasing Solar Energy Adoption Through Roof Detection

Solar AI, a Singapore based startup incubated as a part of ENGIE Factory, collaborated with Omdena, to hyper-scale the deployment of distributed solar and the transition towards 100% renewables by modernizing the way rooftop solar is sold.

Solar energy is a promising and freely available resource for managing the forthcoming energy crisis, without hurting the environment. Unlike conventional fossil fuels, it won’t run out anytime soon.

There’s enough solar energy hitting the Earth every hour to meet all of humanity’s power needs for an entire year.



The problem statement

The rooftop solar assessment process can be time consuming and expensive, taking anywhere between 1 hour to 2 full days to calculate the solar potential of each rooftop. In the solar industry, this has resulted in the cost of sales taking up to 30–40% of total project costs, significantly worsening the unit economics of solar projects.

By automating this process, Solar AI aims to drastically reduce the cost of this process and make this information easily available for both building owners as well as solar energy companies.



The project outcomes


The data

Even the most technically advanced algorithms cannot address or solve a problem without the right data. Having access to data is quite valuable, but having access to data with a learnable structure is the biggest competitive advantage nowadays. Our team of collaborators annotated thousands of rooftops. The consistent determination of the annotators resulted in a perfectly labeled dataset for Supervised Learning algorithms.


The machine learning models

The major task was to detect rooftops in a given image using Computer Vision models. We also had to determine their type/structure such as Flat-roof, Hip-roof, Shed-roof, or any other. Hence, this became an Instance Segmentation problem. The team tried out a number of models such as Mask R-CNN, YOLACT (You Only Look At CoefficienTs), Dectectron2, and more. After training on different batches of annotations as they were delivered, we kept seeing improvement in results. Eventually, the best performing model was selected to go ahead with other tasks.



Now that we had the bounding boxes and mask contours of various rooftops, trapped properly in a data frame, we were ready to start the analysis of individual rooftops. After extracting and zooming into masks of each detected roof, we needed the following attributes:

  • Obstacle Detection
  • Area of the roof (excluding obstacles)
  • Material of the roof
  • Detecting faces of Hip/Shed roof
  • Orientation of individual slopes


For the calculation of a rooftop’s effective area, the area occupied by obstacles has to be subtracted from the whole. So that gives rise to the task of identifying obstacles.

After merging everything into an automated pipeline and many rounds of reviews, evaluation, fixing bugs, and testing — our software was ready to be delivered.

The project falls under the UN´s Sustainable Development Goal 7, which is to ensure access to affordable, reliable, sustainable, and modern energy for all.