Bopeng Zhang

* Gathering model requirements from several units in the bank

* Collecting, cleaning, and analyzing data for model development and testing

* Applying various statistical techniques (e.g. logistic regression, optimization, etc.) to test and select an optimal model for a various data population

* Perform statistical tests to evaluate model power

Extract influential features from large amount of listings on ebay.com using Parsehub in less than a week Build a random forest predictive model in Python on selected features to boost clients sales revenue Analyze impression, click-through rate & sales of campaigns supporting clients executive decision-making

Time series analysis of stock market price trend - Retrieved data from Yahoo Finance and built a database using SQL - Implemented principal component analysis on 10 technical indicators and found 3 most important indicators - Applied linear regression analysis of selected technical indicators to predict price trend in Python - Performed reinforcement learning algorithm to predict price trend in Python - Compared results from two different models and built a recommendation system of price trend Simulation of evacuation of people from Bobby Dodd stadium in Python - Built statistical model of walking speed and destination direction - Applied cellular automata to simulate behavior and interaction between individuals - Analyzed evacuation time and provided solution to alleviating congestion Modeling of reverse electrodialysis system - Solved Navier-Stokes equation numerically in Matlab and conducted stability tests under different boundary conditions - Coupled Nernst-Planck with Navier-Stokes equation to characterize the electrochemical properties of the system - Optimized system by increasing the performance of energy generation based on the simulation results Simulation of supply chain and the effect of demand surge in Python - Implemented event-oriented process to simulation the propagation of product orders in a typical supply chain - Analyzed the effect of a sudden demand surge on the supply chain known as "bull whip" effect Implemented clustering algorithms using K-means, nearest neighbor, and hierarchical in Python and R Applied machine learning algorithms: Classification and Regression Tree, Random Forest, Neural Networks, Support Vector Machine in Python and R

The origin of property enhancement of nanocomposite ion exchange membranes (IEMs) is far from being fully understood. By combining experimental work and computational modeling analysis, we could determine the influence of nanomaterials on the ion transport properties of nanocomposite cation exchange membranes (CEMs). We synthesized and characterized a series of nanocomposite CEMs by using SPPO as polymer materials and silica nanoparticles (NPs) (unsulfonated or sulfonated) as nanomaterials. We found that with the increase of NP loading, measured CEM permselectivity and swelling degree first increased and then decreased. We also found the ion exchange capacity (IEC) and ionic resistance of nanocomposite CEMs tend to be the same, regardless what type of NPs are incorporated into the membrane. Modeling analysis suggests that the change of membrane properties is related to the change in membrane microstructure. With the addition of silica NPs, membrane porosity (volume fraction of intergel phase) increases so that membranes can absorb more water. Also, volume fraction of sulfonated polymer segments increases, which can allow membranes to retain more counterions, causing membrane IEC to increase. By calculating the effective ion diffusion coefficients and membrane tortuosity factors of all the silica-NP-based CEMs synthesized in this study, along with nanocomposite CEMs from previous studies, we conclude that membrane ion transport efficiency tends to increase with the incorporation of nanomaterials. In addition, this paper presents a simulation model, which explains how the membrane property changes upon nanomaterial aggregation; the simulation results are in good agreement with the experimental data. Simulation results indicate that membrane properties are related to nanomaterial number concentration in the membrane matrices; thus, a plateau is reached for membrane ion diffusion coefficients due to the severe influence of aggregation on the increase of nanomaterial real number concentration. The results of this study can provide insight into membrane structure–property relation and contribute to the value of future designs of new nanocomposite IEMs