Hi! This is Qingmengting Wang, you can also call me Pat :)
I am now a Master's student in Data Science at University of San Francisco, with past experiences in Pocket Gems, Baidu and Accenture. I am passionate about applying Data Science techniques to solve real-world challenges.
To know more about me, you can find some info below:
View my resume
View my Linkedin
Machine Learning, Deep Learning, NLP, SQL, A/B testing
Python, C++, Matlab, Spss, Eviews
Spark, BigQuery, PostgreSQL, Flask, ggplot, plotly, AWS, Bootstrap, PyTorch, Git, Keras, TensorFlow
Used several deep learning techniques to work on two natural language processing related projects.
Data Science Intern
Data Analysis Intern
Used exploratory data analysis to detect anomaly for Baidu's advertisting Business.
Conducted oversea-investing strategy research for PetroChina, in which I used several machine learning techniques for some side projects.
A tool that can help you convert your math equation directly to code. This purpose and usage of this project is clearly described in our main page. And you can find our code here.
Will your dog get lost?
Key words: exploratory data analysis, ggplot
This project is aimed to compute the missing rate of different dog’s breeds with DataWorld Animal Center Intake with several exploratory data analysis techniques. We tried to summarize the common features of breeds that are easy to get lost.
This project is a lot fun! After studying a bunch of reasons such as size, age, running speed, we found that the key feature that significantly increases the missing rate of dog is: sterilization, i.e. whether the owner castrates or spays it. For different breeds, unsterilized male or female dogs also have huge difference of missing rate. For example, male huskies have much higher missing rate while female corgis are easier to get lost, from our data.
Patients' heart rate and arterial pressure prediction
Key words: kaggle competition, feature engineering
This project is a kaggle competition that aimed to predict patients' heart rate and arterial pressure. It's a fun competition that relys heavily on feature engineering. To find more, you can see our code here.