Shimin Tu

Logo

Data Analyst, Master of Urban Spatial Analytics @UPenn, Bachelor of Urban Management @Nankai University.

Looking forward to making a more convenient, real and peaceful world using data with humanity.

View My LinkedIn Profile

View My GitHub Profile

Portfolio


Machine Learning Predictions

Worldwide Population Aging and Birth Rate Prediction

There are two main parts in this project, the first part focus on aging level prediction with multiple features extracted from the Health, Nutrition, and Population Statistics dataset and the second part focus on time series prediction for birth rate in 20 years.

House Price Prediction in Boulder County, CO

In this project, house price prediction for Zillow is carried out with multi-resource commercial and socio-economic data. Feature engineering is an important part of the project, which helps achieve a satifying accuracy and generalizability.

Heroin Overdose Prediction in Cincinnati

Faced with the more and more serious heroin overdose in Cincinnati, we trained a model based on Poisson Regression to predict overdose in each area, which is helpful for the Public Health Department to allocate medical resources efficiently.


Dashboards and Reports

Neighborhood Livability Evaluation in Chicago

A series of web-based maps show the evaluation of the safety and convenience of each neighborhood in Chicago based on crime data and local amenities. Residents especially students could choose a reasonable and livable neghborhood to move in, and find local amenities easily!

Neighborhood Livability Report with Power BI

A Power BI version of neighbordhood livability report for Chicago. There are more understandable plots and striking key information!

Heat Islands Exploration in Philadelphia

In this project, we build a prediction model based on Random Forest to predict the temperature if we convert a specific propotion of vacant lands to green space in Philadelphia! We want to use data and model to make the environment better!


Page template forked from evanca