I am currently a post-doctoral researcher at the Department of Earth and Planetary Science, UC Berkeley, United States. Although I am intrigued by all sorts of data science works, my primary research interests are in Geophysics (mostly seismology), and Geodesy. I work on the tools that support data-intensive research and build open-source software tools. I am highly interested in geospatial big data, machine learning, and cloud computing.
Current projects (Not open-source yet)
- Developing a 3D elastic model of deep structure beneath the Yellowstone hotspot (from surface to the Core-Mantle Boundary). This involves:
- Developing and finalizing codes for full-waveform inversion using the Normal-mode perturbation theory
- Implementing full-waveform inversion using the Adjoint method
- Developing and running codes on supercomputing clusters, e.g., Texas Advanced Computing Center (TACC), National Energy Research Scientific Computing Center (NERSC)
- Developing codes that can be parallelized on CPUs and GPUs on any computing facilities (small clusters to supercomputing clusters).
- Visualizing tomographic results by rendering in 2D and 3D.
- Development of a low-cost structural health monitoring system - Quake Structural Integrity System (QSIS). This involves:
- Assembling a single board computing (SBC) system and sensors to build the client system
- Development of the client software to record, process, store, and transmit the data in real-time
- Development of the server software to run on the receiving end of the client software to retrieve the realtime data from multiple clients asynchronously through websocket connection, and post-process data. The server is also responsible for the computation of structural parameters like natural frequencies of the structures, S-wave velocity and damping parameters. It also send alerts to the users by emails or app notifications.
You can check the Github organizations to follow my above two projects:
Open-source projects
- dtwhaclustering: A python package to cluster geodetic time series
- Voice-Plotter: A GUI application to visualize real-time ambient sounds
- System-monitor: A GUI application to get real-time feedback on system runtime statistics
- Miniseed2mat: Convert miniseed data format to MATLAB's native MAT format
- FerryMan: A web application to identify the "comfort" zone of different species based on their Ecological Index
Some Blog Posts
- Using dask Python library to read a huge global earthquake catalog file
- How to plot earthquakes data on a three-dimensional topographic map
- Convert any text to lifelike speech using Amazon Polly
- How to overlay shapefile data on PyGMT Maps
- How to plot Shear-wave splitting measurements using PyGMT
- How effective is the signal denoising using the MATLAB based wavelet analysis
- Empirical Orthogonal Function analysis to inspect the spatial coherency in the geospatial data
- Signal denoising using Fourier Analysis in Python
- Transfer learning applied on the unsplash data using alexnet pretrained network
- PyQt5 Application for visualizing ambient sound in real-time
For more info, visit my blog: earthinversion.com