Upasana Dhar, PhD candidate
Arlington, TX-*****, USA
* ***** ********** ** **********/Data science/Machine learning applications in pre- dictive analysis.
Self-starter, passionate learner, experienced working on diverse domains.Trained in data scrapping, wrangling, feature engineering, training models, cross-validation, reg- ularization, model optimization in Python, scaling in Spark, querying in SQL.
Strong organizational, leadership and communication skills. Mentor and teach courses to students from diverse background. Designed course curriculum with hands-on exercises, which ensued increased student participation and good feedback. Projects (coding in Python, Spark, R, Matlab)
Exposure studies: Temporal trend’s variation of phthalate biomarker concentra- tions, with respect to physical and socio-demographic factors of women, using multivariate regression models in R, Matlab.Paper presented at ISES-ISIAQ 2019, Lithuania.
Health exposure study: Statistical modeling to predict and compare the risk of exposed group living near mines and control group living away from mines. Published chapter in book
Used Random Forest on ApacheSpark to scale big healthcare datasets, predicting customer churn on hospital admissions, for mental or physical health problems, based on diverse socio-demographic parameters.
Developed Image classi cation technique using particle’s form factors, to correlate an- thropogenic input across di erent lake cores.Conference paper at AGU, 2019, San Francisco
Developed multi-layer perceptron neural network for classifying regions of prof- itable oil production, based on di erent input borehole parameters, for UK-CS basin
Experienced in NLP- Developed machine learning model for classi cation of text from humans and from machines. Word Cloud maps for feedback survey data through period of 8 years for temporal variation in Sentiment analysis
Employed K-means clustering to segment neighborhoods based on venue preference of public transits, public places, private restaurants etc, for exploratory BI
Used LSTM neural network for Time-Series forecasting used to predict future stock prices and alpha, beta risk patterns
Spatial imagery analysis and tracing of surface lineaments from Digital elevation models in ArcGIS.Conference paper at IGU 2013
Data Science Professional Certi cate
Databases and SQL for Data Science, Python for Data Science and AI Data Visualization with Python, Data Science Methodology Machine Learning with Python
Advanced Data Science with IBM Specialization
Fundamentals of Scalable Data Science, Applied AI with DeepLearning Advanced Machine Learning and Signal Processing
Advanced SQL for Data Scientists
AWS Fundamentals: Going Cloud-Native
Visual Analytics with Tableau
Research Assistant j Data Analysis j Data visualization (Sept 2015 - Present) Institute: UT Arlington, TX, USA
Project Assistant j Data Analysis j Data visualization (July 2013-June 2015) Institute: CSIR-National Geophysical Research Institute, Hyderabad, India Programming Pro ciency: Python, Apache Spark, R, Matlab, SQL Skills Database management: SQL, PostgreSQL, NoSQL Python packages: Scikit-learn, Numpy, Pandas, Jupyter notebook, Matplotlib, Scipy, Seaborn, Folium
Machine learning frameworks: KNN, Decision Tree, K-means clus- tering, DBSCAN
Deep learning frameworks: Keras, TensorFlow, PyTorch, Feed for- ward networks, CNN, RNN, LSTM, Random Forest, Autoencoder Cloud : IBM, AWS
DataViz: Tableau, NodeRed
Others: ArcGIS, Azure studio, Excel, LaTEX
Internship Drilling Operations and recovery using geophysical techniques, logging and correlation, at Oil India, Assam, India (May, 2012) Co-Curricular Achievements
National-Heritage Scholarship(merit scholarship) at IIT Bombay, India. (2011)
All India Rank 16, in Joint Admission Test for Master’s at IIT, India. (2011)
Departmental Cultural Secretary for 2 years, and bagged the department 1st prize during tenure, IIT Bombay, India. (2013)
Recieved ISGI Grant from Institute of Sustainability and Global Impact,UT Arlington. (2017)
Trained Classical dancer. On-stage performer at various occasions.
Runner’s up in Inter School Kho-Kho Championship, represented school team for 3 years. Publications and Conference Presentations
Quantifying anthropogenic inputs in contaminated sediments based on particles’s form factors using machine learning techniques on CCSEM data, IGC 2020, Delhi, India
Temporal trends of phthalate exposures in California’s pregnant women: Comparison with NHANES data, ISES-ISIAQ 2019, Kaunas, Lithuania
Other publication include 3 articles, 2 conference publications and 1 co-authored chapter in Book For more information check Google Scholar