Mikael Brunila's Projects
Data and models to extract toponyms and spatio-temporal entities from text data.
Python scripts written during the course "Programming in Social Science" at the University of Helsinki.
Presentation in the working group on 'The Digitalization of Societies & Methods' at the Finnish Sociology Days. Scraping & mapping data from Twitter and Facebook with R, Python, Open Refine & CartoDB.
This is my final project for a GIS class at Columbia, exploring the relationship between tweeting and gentrification.
minä minä minä
GitHub Pages
Analytics web app in Shiny using data from Uber and MTA.
Analysis and Visualization of Interconnected Multilayer Networks
Information theory in NLP for people without a Math background
Text summariser written using Keras and TensorFlow for the NLP class of Professor Kathy McKeown at Columbia University.
A very rudimentary homework assignment in using Keras for Sentiment Analysis. Part of the NLP class at the Columbia DSI.
Some Python scripts for classifying tweets as either Democratic or Republican. Homework assignment in the NLP class by Kathy Mckeown at Columbia Uni.
Homework assignments for the data visualization class at the Columbia University QMSS program.
course material
Lab assignments for the Social Network Analysis class at the Columbia QMSS program in the spring of 2018.
Add CRF or LSTM+CRF for huggingface transformers bert to perform better on NER task. It is very simple to use and very convenient to customize
Tietokantasovellus-kurssin aloituspaketti
This data mining project for Columbia uses LDA topic modelling to map tweets in New York.
Calculates Word Mover's Distance Insanely Fast
A Python implementation of Word Mover's Distance that decomposes document level WMD into word level WMD for interpretable sociocultural NLP.