Skip to content

Latest commit

 

History

History
69 lines (63 loc) · 10.1 KB

README.md

File metadata and controls

69 lines (63 loc) · 10.1 KB
Raw

An incomplete collection of example coding projects & code snippets

Feel free to also check out my blog that has and will discuss some of my past and future projects as well as provide guides regarding data science and networking.

The Tech stack for all of these project (among others):

Visual Studio Code Git GitHub Docker
Python JavaScript Rust
HTML5 CSS3 Flask
Postgres MongoDB
NumPy Pandas Matplotlib Plotly
Scipy scikit-learn TensorFlow Keras

Apps and app-likes

  • Berlin Bürgerservice Bot: JS Webscraping app that scans Berlin's municipality appointment service website for available timeslots of Berlin's Administrative Office for Citizens.
  • Chat-GPT CLI Integration: A CLI integration of Chat-GPT3.5 written in Rust.
  • Color wheel: Color-wheel generator for isoluminant colors in HSV, Cie-L*ab and Xyz space ready to be applied to PsychoPy.
  • Pokerstars Reader: App that provides live table and player statistics when playing Poker on a PokerStars table.
  • RPi: Face recognition: App that allows Raspberry Pis to detect and recognize faces using one's own pretrained face recognition model.
  • RPi: LCD Moisture Meter: This C++ app enables Raspberry Pi Pico microcontrollers to display soil humidity levels, as measured by capacitive sensors, on an LCD screen.
  • RPi: Motion detection: App that allows Raspberry Pis to detect motion based on PIR sensor input.
  • Tennis DB API: A simple API written in Rust meant to retrieve data from the Tennis subreddit PostgreSQL database that was created as part of the Sentimental(al) Reddit Slack Bot project.

Modules / Wrapper

  • PrettyShortML: Scikit-learn wrapper Python module for faster ML model pipelining, better plotting and more.
  • Scipy statistical utils: Set of functions for calculating signal-detection theory metrics (a', d', Beta and criterion), within-subject confidence intervals with Morrey's (2005) correction applied and cluster-based permutations implementing methods described in Maris & Oostenveld (2007).

Data Science

  • Anomaly Detection: Finding abnormally low salaries and water pump failures: Approaches to detect anomalies in static and time series datasets using KMeans clustering and Isolation Forest modelling in scikit-learn.
  • Classification: Identifying Animals from Pictures: Implementation of the MobileNetV2 deep neural network to classify animal (races) from pictures using Tensorflow / Keras.
  • Classification: the Titanic dataset: A Kaggle Competition dataset to predict the survival of passengers from the famous Titanic cruise ship using classification models in scikit-learn.
  • Dashboard (Metabase): Northwind dataset: Interactive dashboard for analyzing the Northwind database using AWS RDS & EC2, PostgreSQL and Metabase.
  • Dashboard (Streamlit): HR dataset: A simple Streamlit dashboard visualizing an HR dataset using Streamlit and Plotly.
  • Deep Learning Objects: webcam images: An artificial and convolutional neural network that classifies object categories in webcam images using Tensorflow / Keras.
  • Deep Learning Objects: the MNIST dataset: An artificial and convolutional neural network that classifies handwritten digits from images of the MNIST dataset using Tensorflow / Keras.
  • fMRI NRoST: Data analysis project of a functional magnetic resonance imaging (fMRI) experiment for a scientific peer-reviewed publication.
  • Natural Language Lyricizer: Implementation of a natural language processing model to predict music artist from song lyrics alone using webscraping and Bag of Words modelling.
  • Sentiment(al) Reddit Slack Bot: A Slack bot that provides live updates on positive and negative postings regarding the topic of Data Science using Docker, API interactions, MongoDB, PostgreSQL and VADER-sentiment analysis.
  • Object-tracking in sports: A computer vision processing pipeline that tracks tennis players and the tennis ball in videos of tennis matches using OpenCV and object detection with YoloV8 and TrackNetV2 in Tensorflow (Keras and Pytorch).
  • Supermarket Markov Simulation: Visualization of supermarket customer behavior based on Monte Carlo Markov Chain Simulations.
  • Regression: the Bikeshare dataset: A Kaggle competition dataset to predict bike rentals from the BikeShare dataset using regression models in scikit-learn.
  • Temperature forecast for Berlin-Tempelhof: Visualization of a Time-series analysis on temperature data, i.e. a temperature forecase, for Berlin-Tempelhof using AR/SARIMAX modelling and interactive plotting in Plotly.
  • Unsupervised Movie Recommender: Web app that recommends API-requested movies based on unsupervised learning modelling (Neighborhood-based Collaborative Filtering and Collaborative Filtering with Non-Negative Matrix Factorization) deployed with Flask.

Science