project-sebi Goto Github PK

• 👋 Hi, I’m @Project-Sebi
• I’m a Bachelor student of Philosophy at KCL
• 👀 In terms of programming, I'm currently interested in data science and machine learning. I am comfortable wit Python less so with R and I also have experience with Java, but from a few years back.
• 💞️ I’m looking to collaborate on Machine Learning projects and learn ML methods thoroughly.
• 📫 How to reach me: [email protected]

COMPLETED PROJECTS:

Causal Inference and Machine Learning for Politics - Module at Humboldt University. All programming is in R. For Each week one problem set. Problem sets are uploaded here at my github.

• First half of the module focused on Causal Inference methods such as Difference-in-Differences, Instrumental Variables, Interaction Effects, Reggression Discontinuity Design.

• Second half of the module focused on Machine Learning methods such as dimension reduction for OLS reggression (Lasso and Ridge Reggressions), Logistic Reggression, Tree-based methods (bagging, boosting, random forests), Support Vector Machines (theory only), Deep Learning (Theory only).

• PS 1 is an Introduction to Causal Inference framework using 2012 Field Experiments: Design, Analysis, and Interpretation by Gerber and Green, Chapter 1 exercises

• PS 2 is a general introduction to R using https://rafalab.github.io/dsbook/

• PS 3 is introduction to R's tidyverse using Chapter 4 of https://rafalab.github.io/dsbook/ and some exercises from Chapter 2 Gerber Green 2012

• PS 4 is a completion of exercise from Chaper 7 ie. ggplot() https://rafalab.github.io/dsbook/ and solving Instrumental Variables problem using AER package for R

• PS 5 is a data cleaning and a simplified replication of 2011 Regression Discontinuity Design paper Incumbency Advantage in Brazil: Evidence from Municipal Mayor Elections by Rocio Titiunik

• PS 6 is a simplified replication of a 2012 Diff-in-Diff paper Commodity Price Shocks and Civil Conflict: Evidence from Colombia by Dube and Vargas

• PS 7 uses Matching method between similar treated and control subjects

• PS 8 uses interaction effects and Matching methods

• PS 9 was introduction to Machine Learning (bias, variance, K-nearest neighbors)

• PS 10 was regression and classification

• PS 11 looked at sparse and flexible regressions (lasso, ridge, smoothing by derivation)

• PS 12 was introduction to decision trees

CodeAcademy course, Exploratory Data Analysis in Python pandas, numpy, matplotlib, seaborn corresponding files: Data Cleaning in Python, Data Visualization in Python, Aggregate Functions in Pandas

UNDISCLOSED PROJECTS:

• Non-neural networks NLP for Prague start-up MAMA.AI
• Building conversational trees
• Automating work of call centres
• I cannot reveal the code due to Non-Disclosure Agreement