Data Science Intern
Current• Built a time series forecasting model for measurements of electric power consumption in households with the potential of preventing overloading and allowing efficient energy storage, the model achieved a MAPE of 4.2% (96.8% accuracy).• Built an AI model using Python to label satellite image chips with different classes of land cover/land use • It was a multi-labeled classification problem with 17 categories, achieving an F2_score of 85%.• Led a team of 20 Interns to analyze and build a regression model using Time Series data to identify market trends and predict the future adjusted closing price of Gold ETF across a given period in the future. It has the potential to help investors - government or private - to decide when to buy or sell the commodity, as any fractional change in its price may result in a considerable profit or loss. It recorded an R-Squared score of 98.8%.• Build a model to classify and predict the quality metric (QAScore) ecological footprint data for different countries.• Successfully developed a multivariate regression model to study the effect of eight variables on residential buildings' heating and cooling loads.
