Jing Ren Email & Phone Number

Irving, Texas, Us

• Conducted association rule mining to detect frequently bundled events. Carried out preceding event analysis and mean time between failure (MTBF) analysis to understand data distribution.• Created features based on most recent occurrences along with stratification, rebalancing, and 3-fold CV.• Built machine learning models (logistic regression, random forest, XGBoost, SVM) and time sequence models (simpleRNN, GRU, and LSTM) to alert critical equipment failures 7 days in advance and reduce unexpected machine downtime.

Northbrook, Il, Us

• Build recommender system with matrix factorization algorithm based on product reviews, browsing history, and transactions. Perform market basket analysis to increase attachment rates, sales and average order value.• Provide insights on how to merchandise products differently to capitalize on in-store and online shopping preferences.• Identify localization trends in product categories and customer segments to improve BOPS/BOSS offerings.

Atlanta, Ga, Us

• Perform EDA to quantify package path, sort characteristics, and error occurrences in sort facilities. Investigate potential causes for extremely high times and the correlation with error rate with over 2 million package sorting datasets.• Identified important features, analyzed their effects on package sort time through feature importance and ALE/PD plots.• Build ensemble models (random forest, neural networks, XGBoost) by feature engineering and parameters tuning with Python to predict package sort error and sort time across facilities.• Create a real-time dashboard with Tableau for managers to monitor sort metrics and leverage them to make business decisions.• Research best practices for scaling the model to new sites to ensure incremental updates while new data are collected.

Denver, Colorado, Us

• Investigated whether damaged houses are qualified for retrofitting after natural disasters using deep learning models.• Used OpenCV (Edge Detection, Histogram of Oriented Gradients, etc.) to extract the most significant features based on 10,000 configurations of building facades.• Constructed XGBoost classifier based on the training dataset of generated images with labels GO’s (houses that qualified for retrofitting) and No GO’s, to which successfully applied to real house images with approx. 70% accuracy.

• Identified traffic patterns in time series with tracking data such as click counts, page views, and visitor sessions.• Created data visualizations using Tableau and Python to illustrate user demographics. • Proposed website redesign ideas to increase website traffic through analyzing user behavior and preference.

Geometry of Traffic• Investigate parking conditions in San Francisco using data from geotagged parking behavior data and public API. • Apply mathematical algorithms to the dataset of 19 million lines and clean data using Python.• Create visualizations using Circle Map and analyzed parking behavior like pricing adjustment and occupancy change rate.• Use low-rank factorization clustering to identify lot distribution and correlate with real estates and populations. Surfaces Moving Under Hamiltonian Flows• Constructed an algorithm that optimized Going-Jaeschke’s gradient descent algorithm against Haim-Hislev’s formula using Python and calculated the Ekeland-Hofer-Zehnder (EHZ) property for surface movement under Hamiltonian Flows.• Estimated EHZ capacity of convex polytopes and tested Viterbo’s conjectures on large sets of random polytopes to examine the relationship between EHZ capacity with flow volume.• Refined algorithm efficiency on iteration by collecting data and solving corner cases, which increased polytope capacity and lowered overall algorithm running time.

Gb

• Assessed client’s control risks through reviewing 300+ sampled transactions in revenue, disbursement and payroll cycles for proper approval, recognition and record keeping and examined variations from the standards. • Tested cash accounts from 31 branch offices by reviewing bank reconciliations, statements, outstanding check and deposit in transit supports, confirmed account balances with banks, and investigated variances. • Projected profit margins based on past data and changes in key business drivers and performed variance analysis.

• Forecasted market trends, revenue streams and operation expenses to develop 2017 budget for department. • Evaluated detailed program bidding offers; calculated internal rate of returns using free cash flow methods.

Master Of Science - Ms, Analytics

Bachelor Of Science - Bs, Mathematics And Statistics, Minor In Cs