Name: Tyler Burns
Type: User
Company: Burns Life Sciences Consulting, GmbH
Bio: I'm a classically trained biologist who learned programming and statistics thereafter. I am now a hybrid who solves problems utilizing both skill sets.
Location: Berlin, Germany
Blog: www.tylerjburns.com
Tyler Burns's Projects
Type a question. Get the passages in Meditations by Marcus Aurelius that are most relevant.
Some helper functions to visualize the stability of a clustering method for your single-cell data.
Convolution-based cellular automata, with kernel adjustment leading to new rules.
A brief exploration of different ways to transform CyTOF data.
How to analyze one CyTOF fcs file.
An interrogation of the U-Matrix as a visualization of SOM-clustered CyTOF data.
Adding noise to data being used as input for dimension reduction tools
Run t-SNE or UMAP 100 times or more. How does the island placement change between runs?
View a spectrum of dimensionality reduction embeddings for your single-cell data ranging from t-SNE like to UMAP like
View several hundred search results at a time with a topic map.
Convert your journal into a mappable, searchable vector space
A wrapper around the sleepwalk package that allows for interactive knn visualizations of an embedding.
R package that allows for visualization of the KNN of an embedding compared to the KNN of the original high-dimensional data.
Using entropy to determine relative efficiency of written languages. I was trying to figure out if English or German is more efficient.
Make KNN-based identity comparisons between different manifolds (eg. original space vs t-SNE space). For a comparable project using distance comparisons see https://anders-biostat.github.io/sleepwalk/ by by Svetlana Ovchinnikova and Simon Anders
Scrapes bio/chem/medrxiv preprints and places them into a Dash table that you can filter.
Generates a sortable table of biorxiv and medrxiv twitter handle metadata.
Programmatically pull pubmed search results for predictive analytics.
Make a context map of your RSS feed
Group CyTOF data into overlapping k-nearest neighborhoods for enhanced single-cell visualizations.
Exploring dimension-reduced embeddings
Create searchable and sortable tables of the tweet histories of your favorite twitter users.
Mining automated paper-posting bots from twitter, especially useful to determine which pre-prints are being talked about.
Scrape tweets from any number of users, embed the tweets, do dimension reduction and clustering, and visualize them in an interactive interface.
My early solution to running UMAP on your flow and mass cytometry data and compare it with t-SNE. Now I use the umap package in CRAN.