The well log data are consists of physical properties of measured rock or lithofacies on certain depth in earth. These physical properties are unique for each lithofacies. So, I used to run some clustering technique such as K Means Clustering and HDBSCAN to cluster some group with the same physical properties and tried to interpret the lithofacies group from this clustering result. Also I used dimension reduction technique such as Principal Component Analysis (PCA) and t-SNE to simplify the process and increase the accuracy.
Because the interpretation of well log is the main objectives of petrophysicist, I tried to develop a desktop application based on this technique, so it can help the petrophysicist interpret the well log data more accurate and more efficient. I used PyQt5 as the backend of this application and it completely run on Python.
Skills: Python, Machine Learning, Numpy, Lasio, Pandas, Matplotlib, Seaborn, Scikit-Learn, PyQt5