The Impact of Data Analysis on Solar Power: A Study on how data analytics improves an Individual's solar solution in South Africa

Abstract

This paper explored the feasibility of solar power systems for individual use. With the increase in load shedding and the lack of reliable electricity in South Africa, many households were looking into alternative sources of electricity. One of the main alternative sources was solar power. Many different factors could affect the amount of usable electricity that solar power systems can produce, including weather and the position of the panels. With the idea of solar power came many questions, with the biggest being whether solar power was feasible for individual households. With the help of data analytics, this question was answered by running a few simulations to determine if it was worth installing solar and, if it was, what the best way to set it up was. This was achieved by using a Python library that ran many different calculations and simulated the outcome using the given data. The library took in many different factors, including the coordinates, to determine the main weather patterns, the tilt of the panels, the number of panels, as well as start and end dates. This meant that data could be simulated for years to get the most accurate outcome. With all of this data, it could be used to determine if it would be feasible for an individual to use solar. This could be compared to a person's normal electricity usage to see if it would 1) generate enough electricity for the household and 2) what the optimal set up would be. This was done by creating a machine learning algorithm that took in all the information and generated the results for an individual household. It was found that while it might be able to sustain some households, enough power is generated to make a substantial difference.