Ordered newest to oldest
Click the title of any of the streamlit applications to open them in another window.
In this application, I loaded in data from various government sources about the population and number of charging stations of different PLZ (postal codes) in Berlin. A heatmap overlay of the map of Berlin, subdivided by PLZ is then presented with a filter for number of inhabitants, and a filter for number of charging stations. In the second part of the project, I worked with a Team to implement a demand estimation layer for the map, and a suggestions box which can be filtered by postal code.
In this application, I used publicly available IRS data here and processed it to find the median income of any zipcode in the United States that is described by the data. The information used to do this was the number of tax returns in each tax bracket and the bracket values themselves. I made the assumption that the income in each bracket rises linearly and therefore used linear interpolation to approximate the median.
Here I showcase a few GitHub projects on machine learning and data science.
In this project, I built a feed forward neural network capable of classification and regression all the way from the matrix multiplications. It was written in python using the NumPy library so that we could deeply understand the mathematics behind neural networks.
Check it OutIn this group project, I led a team to conduct an analysis of predictors for housing prices using R and techniques from multivariate statistics.
Check it OutThis project was concerning the prediction of stellar parameters given light curve data through the use of deep learning techniques. I worked with two other seniors in computer science and data science to experiment with various techniques and produce the following writeup.
Check it Out