uYilo Electric Mobility ProgrammeData Science Intern
Jul. 2021 - Aug. 2021• uYilo faced a challenge of not understanding where in tests, the electric systems failed. I have been selected out of cohort of 20 students to work with the company to analyze the system data and understand what is causing the systems to fail and to predict when future failures would occur. • Used Python Libraries such as Numpy, Pandas, Scikit-learn, Matplotlib and Seaborn to conducted exploratory data analysis on renewable energy system data where the system experienced a trip and stopped working to better understand which components experienced failure and which system variables are highly correlated. • Collaborated with 3 colleagues to find possible outliers on highly correlated system variables and observed early signs of system failure and created visuals through Plotly, Seaborn and Matplotlib libraries to explain the findings to both technical and non-technical audience, and provided insights on steps the company should take to monitor early system failures. • Ran different regression, classification models to find the best model for the dataset and concluded random forest regression with an 87% accuracy score through confusion matrix which the company can implement into production for Big Data.