|
I do interdisciplinary studies of mental health and data mining. My research interests lie in creating interactive mental health intervention tools that can better serve underserved population at a low cost. I obtained my bachelor degree, major in general psychology and minor in applied statistics, at Univeristy of Michigan, Ann Arbor, and received my master in Statistics at Columbia University. In my leisure time, I enjoy basketball, hiking, and board game. I am a 4th year PhD in personal health informatics @ Northeastern University in Boston, and my advisor is Prof. Stephen Intille. If you are interested in me, don't hesitate to send me an email! |
|
|
|
 |
Automated Semantic Enrichment of Trajectory with OpenStreetMap
This study explores the potential of using free open-source OpenStreetMap(OSM) geodatabases in semantic enrichment of individual’s frequented places. We created a three-level hiarchical taxonomy of place type labels, and a data processing pipeline from segmenting trajectory data to labeling visited places with OSM semantic tags. For a dataset with 93 people's 1-year location data, we successfully labeled 81.9% identified places and 60.3% of labeled places are matched with the groundtruth. The code base of curating OSM data and semantic enrichment will be open-source for other researchers to use. |
 |
Contextual Biases in Microinteraction Ecological Momentary Assessment Non-response
Microinteraction EMA (μEMA), is a type of EMA where each survey is only one single question that can be answered with a glanceable microinteraction on a smartwatch. Prior work shows that even when μEMA interrupts far more frequently than smartphone-EMA, μEMA yields higher response rates with lower burden. We examined the contextual biases associated with non-response of μEMA prompts on a smartwatch. Based on prior work on EMA non-response and smartwatch use, we identified 10 potential contextual biases from three categories: temporal (time of the day, parts of waking day, day of the week, and days in study), device use (screen state, charging status, battery mode, and phone usage), and activity (wrist motion and location).
|
 |
MixWILD: GUI-based Desktop Application for Intensive Longitudinal Data Analysis
MixWILD (Also Mixed model analysis with Intensive Longitudinal Data) is a desktop GUI-based application for examining the effects of variance and slope of time-varying variables in intensive longitudinal data, especially the ones collected using ecological momentary assessments. In this ongoing project, we keep working on new statistical modeling functionalities and GUI features for health science researchers to employ complex multi-level modeling with just a few clicks. Github Website |
|
Personality Impressions of YouTubers Based on Their Trailers and Channel Sites: A Lens Model Approach
This study is to investigate the discrepancy between the actual personality of creators of the popular YouTube channels, the presenting personality that the YouTube channel intend to communicate with their audiences, and the personality impressions to the common audiences. A lens model analysis will be used to find what cues observers depend on when forming personality impressions of YouTubers. Consensus and accuracy of personality impressions will also be examined compared to the actual and presenting personality. Gabriella M. Harari |
|
|
 |
Maths and Physics Exercises Text Classification with Concept Map
Built innovated ensemble classifiers to predict the major thread of concepts of maths and physics exercise texts, given the concept maps with a multiple-layer hierarchy. The classifier features small size training set, automatic tuning with new data and good scability to enlarge concept maps. |
 |
Transportation Departure and Arrival Delay Prediction
Web scraped weather and airport information and combined unstructured data sources including departure and destination information, GPS location, customer comments to make predictions of transportation delay. Stacking models and careful feature engineering were applied and best performance were achieved through fine-tuned neural network. |
 |
Coupon Recommendation System for Retailers (Design)
Designed collaborative filtering recommendation system for retailers to distribute coupon prize in a lottery game to attract customers to repurchase the products. Given no knowledge of coupon attributes, the collaborative filtering system integerates the customer demographics and historical prize redemption records and adapts to individual hidden preferences through constantly referring to the coupon redemption choices of similar customers. |
|
|