My Data Projects
Explore my data analysis and machine learning projects using the Pandas module!
Exoplanet Detection with Fourier Transform and Oversampling
To better capture periodic patterns in the stellar light curves, I implemented a Fast Fourier Transform (FFT). This mathematical technique converts time-series data into the frequency domain, highlighting hidden features and improving signal detection. I created a custom fast_fourier_transform() function and applied it across both the training and testing datasets.
Data Exploration with Pandas/Slicing and Visualization
In this project, I expanded on data visualization (boxplots) by learning how to slice Pandas DataFrames to extract specific rows, columns, and subsets. Slicing is crucial for cleaning, filtering, and preparing datasets for analysis. This work highlights practical skills in data manipulation and preprocessing, forming the foundation for deeper data analysis and machine learning.
Meteorite Landings - Histogram
In this project, I analyzed the Meteorite Landings dataset by creating histograms to visualize the distribution of meteorite masses and locations. Histograms helped reveal patterns such as the frequency of small vs. large meteorite landings and provided insights into the skewed distribution of the data. This project demonstrates skills in data visualization, exploratory data analysis (EDA), and interpreting real-world datasets using Python.
Pulsar Star Classification
Pulsars are rare neutron stars that emit detectable radio signals. Using the HTRU2 dataset, I trained an XGBoost Classifier on 8 signal-based features (statistics from pulse profiles and DM-SNR curves). Since pulsar examples are a minority class, I addressed the class imbalance to improve detection accuracy.