(Acquired by Foursquare)

As a Senior Data Scientist at DataKind, I have accomplished the following:  •Built a Random Forest classifier to detect wheat lands in Montana and then transferred and tuned it in Ethiopia•Employed temporal cluster analysis by means of feature engineering to identify crop phenology in Ethiopia•Created a visualization tool to explore temporal precipitation, also crop types and their greenness, using Tableau•Developed a crash index based on NYC311 complaints using Kernel Density Estimation, for NYC Vision Zero Project•Utilized the NYC CitiBike database to create the normalized index for spatio-temporal analysis of biking accidents •Analyzed spatio-temporal patterns of mining activities in Peru and DRC to identify historical gold rushes in the regions •Scoped and managed several data-driven projects in a variety of sectors, such as Poverty, Education and Immigration• Proposed and served as technical lead on a 5-pronged project that uses satellite imagery and transfer learning technology for poverty alleviation, becoming a top-10 front-runner of the on-going IBM AI XPRIZE Competition

As a Data Scientist, I have accomplished the following: • Mined the records of 10 million NYC311 calls and performed geospatial and temporal aggregation of all data points to census tract using Panda & Arcpy• Created web-interactive visualizations of flood-related NYC311 calls using PostgreSQL & OpenLayers JavaScript API • Built a vulnerability map for the sewer system using Open Data to be presented to NYC Department of City Planning• Mined images taken at 2-minute intervals from 400 NYC traffic cameras and developed a 6-month meta-database• Correlated the rain signal with rain gauge data collected from the NOAA online database using SciPy• Mentored Master’s students who quantified PM2.5 exposure in New York City in the following ways: a) framing the problems of a general lack of spatial resolution of PM2.5 measurements and temporal variability of crowd movement; b) guiding students on exploration of different datasets, including census, infrastructure, and meteorological records; c) supervising the implementation of visualization techniques to understand the correlation between air quality, metrological factors, and human activities, such as Scatterplot Matrices; d) assisting students in the creation of a spatiotemporal PM2.5 exposure model including support vector machine regression

As an Insight Data Science Fellow, I have accomplished the following: • Developed a prototype comparing the operations & finances of shared rides to LaGuardia vs. JFK Airport for a startup• Designed an A/B testing experiment to target persuadable voters and implemented a predictive model to estimate the turn out rate for each zip code in Los Angeles County for the California primary election of 2016 • Created an interactive web-application to visualize and explore factors relating to patients' length of stay in hospital using D3 [http://signal.nyc]• Developed a database using NY State SPARCS Dataset of 12 million patient records from 2009-2013 in MySQL• Employed Linear, Random Forest, and Gradient Boosting Regression Analyses to predict patients’ length of stay based on characteristics such as diagnosis, severity of illness, age group, and gender• For Insight program preparation, mined and mapped the 50,000 geo-referenced tweets made during a one-week period for lower Manhattan using Matplotlib Basemap Toolkit as well as Geopandas and Shapely packages

As a Research Associate and Projects Liaison, I have accomplished the following: • Initiated and Managed the project [NYC Flooding Impact Using 311 Call Records] as part of New York State Resiliency Institute for Storms and Emergencies Phase II Proposal • Initiated and Managed the project [Department of Transportation Cameras as Rain Gauges] for incorporation in the NOAA-CREST Institute’s 5-Year Funding Proposal• Mentored a PhD candidate who is developing a landslide hazard map for the continental United States using remote sensing techniques in the following ways: a) supervising the use of satellite data including Advanced Thermal Emission and Reflection Global Digital Elevation Map (ASTER-GDEM) and Soil Moisture Active Passive (SMAP) monitoring for development of a global shallow landslide susceptibility and hazard map utilizing logistic regression analysis; b) introducing the idea of using a self-designed Buffer Threshold analysis method to improve ability to determine slopes at highest risk of landslides; c) critiquing all preliminary drafts for future journal submission

As Research Assistant, I accomplished the following:• Evaluated the resiliency of the NYC subway infrastructure to extreme events using MTA ridership data• Developed a binary image processing technique to digitize the NYC sewer network in MATLAB• Employed colorimetry to identify night-light sources in NYC to approximate city energy consumption

As Graduate Student Researcher, I accomplished the following:• Applied imaging techniques and concept of color analysis to calculate the volume of oil in soil using MATLAB • Innovated the 3D Carving Technique to estimate the shape of the oil plume inside soil • Developed a 3D Exhaustive Block Matching Algorithm to track color tracers within transparent media

I instructed students in the following courses:• Fluid Mechanics Laboratory 2013-2014• Soil Mechanics Laboratory 2011-2013• Environmental Engineering Laboratory 2010-2011