Nandini Kollati

Data professional pursuing Master's in Information Technology (AI) at Macquarie University with prior experience in utility sector analytics. Strong analytical mindset with passion for machine learning and statistical modeling. Committed to transforming complex data into actionable insights through innovative visualization techniques. Combines troubleshooting abilities with continuous learning approach to stay current with emerging AI technologies. Collaborative team player focused on delivering data-driven solutions that support strategic decision-making.

• Programming for Everybody (Getting Started with Python), University of Michigan, Coursera
• Exploratory Data Analysis With Python and Pandas, Coursera Project Network, Coursera
• Basic Statistics in Python (ANOVA), Coursera Project Network, Coursera
• Introduction to Programming with MATLAB, Vanderbilt University, Coursera

UGVCL Power Utility, Demand Side Management (DSM): This module aimed to efficiently manage and  balance the electrical demand and supply during peak load consumption periods.

• Load Forecasting Expertise: I engaged in load forecasting across various timeframes, spanning very short term, short-term, mid-term, and long-term predictions. These forecasts were instrumental in proactively managing and responding to the dynamic energy demands.
• Model Diversity: To ensure the highest forecast accuracy, I harnessed a blend of cutting-edge forecasting models. This included the application of FB Prophet, Multi-Linear Regression, ARIMA, and LSTM models. These techniques were implemented in both Python and R languages, showcasing my versatility in programming.
• Data Management Excellence: My role required an intricate approach to data management. In the case of load forecasting, I directly interfaced with databases, fetched data through queries, and seamlessly updated the database with the forecasted values. This end-to-end process was completed using Python, enhancing the efficiency of the DSM.

SLMS (Street Light Management System): The project extends its capabilities to real time monitoring of thefts, tampers, and faults, promoting swift and informed responses to potential  disruptions. Notably, the system facilitates remote connectivity for efficient facility maintenance,  allowing for remote connect/disconnect operations

• Theft Detection: Reducing losses due to theft.
• Exceptions: Improving energy efficiency.
• Blown bulb prediction: minimizing disruptions caused by blown bulbs.

Power BI Dashboards & Reports: This Power BI initiative showcased my proficiency in  data visualization, geospatial analysis, and DAX functions, contributing to more informed decision-making  processes within the organization. The resulting dashboards provided a powerful tool for stakeholders to  gain actionable insights and navigate complex data landscapes effectively.

• Data Source Integration: Connected seamlessly to SQL Server, extracting and transforming data for meaningful analysis within Power BI.
• Visual Representation: Developed interactive dashboards featuring a variety of visual elements, including Funnels, Card, Gauge, and Slicers. These components provided a comprehensive view of critical metrics and KPIs.
• Organizational Hierarchy: Illustrated the office hierarchy through intuitive visuals, facilitating a clear understanding of the organizational structure.
• Meter Communication Analysis: Distinguished between communicated and non-communicated meters, offering a valuable insight into the status of communication infrastructure.
• Infrastructure Overview: Showcased the total count of sub-stations, feeders, Distribution Transformers (DTRs), and individual meters, providing a holistic overview of the electrical infrastructure.
• Geospatial Visualization: Leveraged longitude and latitude data to map the physical locations of meters, enhancing spatial awareness and facilitating strategic decision-making.
• DAX Functions Implementation: Implemented advanced Data Analysis Expressions (DAX) functions within Power BI to create new columns and measures. These functions added depth to the analysis, enabling the extraction of meaningful insights from the data.

PREDICTION OF BENIGN AND MALIGNANT TUMOURS IN LUNG CANCER

• Lung Tumor detection using deep learning techniques.
• Captured Image dataset from Kaggle.
• Implemented ConvNet, VGG16 & ResNet50 Algorithms.