Alex Ruan's repositories
ZovcIfzm
Personal GitHub description
react-native-templates
Templates to quickly build an app off React Native
ItsDinnerTime
A mobile app for groups sharing an apartment detailing the menu and who's cooking for the next couple days. Currently being built with a React-Native front end and a Firestore backend. Has some legacy flask/sqllite/jinja2 code. Production is halted to work on DiscoverOHS.
legacy-personal-website
Deprecated personal website- outdated styling/info. Plan to completely rewrite in the future.
ExploreOHS
An mobile app built on React Native allowing users to explore Okemos High School using a network of traversable pictures through gestures.
chem-club-website
Website for the OHS Chemistry Club. Built with HTML/CSS & Bootstrap, hosted on Google Cloud App Engine.
react-stripe
minimal working react app integrated with stripe
BasicClickerApp
An example app for basic react native functionalities
react-native-stripe
Working payment processing for a react-native mobile app using Stripe
minimal-flask-firebase
A minimal example app that shows you how to connect to and call firebase in python from a flask app.
Wound-Analysis-Frontend
A system for the capture, automatic area measurement, and analysis of wounds in diabetes utilizing computer vision with OpenCV integrated with a Flask API for image analysis and a React web interface. This repository consists of the frontend functionality.
Wound-Analysis-Backend
A system for the capture, automatic area measurement, and analysis of wounds in diabetes utilizing computer vision with OpenCV integrated with a Flask API for image analysis and a React web interface. This repository consists of the backend functionality.
EELS-simulator
A simulator modeling the evolution and energy loss of the phase space of a high-intensity electron beam as it travels through a transmission electron microscope with adaptive optics.
minimal-react-firestore
minimal app for connecting react with firebase 9
ctflearn-solutions
solutions for ctflearn problems
Wound-Analysis-Auxiliary
Repository for additional code not required for live server deployment but necessary for preprocessing.
ecs-fargate-tasks-fatlambda
Template code to create tasks as Amazon ECS- Fargate Containers and run them using using event-triggered Lambda functions
Wound-Analysis-ML
A system for the capture, automatic area measurement, and analysis of wounds in diabetes utilizing computer vision with OpenCV integrated with a Flask API for image analysis and a React web interface. This repository consists of the ml functionality.
Kimball-ATVG-pathway-enrichment-analysis
Pathway analysis of differential gene expression for (Ly6C Hi - Ly6C Lo) cell states in DIO vs ND mice. We find significant upregulation in the PGAM1 and downregulation in the CHCHD2 gene.
one-shot-learning-with-siamese-networks
Investigations on one-shot learning with siamese networks based on code from https://towardsdatascience.com/one-shot-learning-with-siamese-networks-using-keras-17f34e75bb3d
Kimball-ATVG-outlier-analysis
Outlier analysis of log-fold-change of gene expression between DIO and ND mice during low and high Ly6C phases with extreme value analysis and DBScan.
InTheLoop
To help those with muted groups chats still hear about interesting texts, InTheLoop, will notify you when user-defined flags are hit in the chat.
minimal-react-context-app
App with minimally working react-context
tensorflow-rnn-predict-fuel-efficiency
Using an DNN to predict fuel efficiency of 1960-1970's cars from the Auto-MPG dataset.
DeepHashing
Image hashing is the one-way process of converting an image into a binary hash such that similar images have similar hashes. This has promising applications in speeding up approximate nearest neighbor search when trying to retrieve similar images from a database as well as in security for verifying an image hasn’t been perceptually modified. Utilizing deep learning, we implement a model that learns these binary hashes under three primary constraints. First, we minimize the loss in information between the continuous model output and the quantized binary hash. Second, we make sure the binary values are distributed evenly on each bit. Third, we ensure different bits are as independent as possible through a relaxed orthogonality constraint on each fully connected layer of the model. In addition, we implement a variant of the same model that takes advantage of training data labeled for classification tasks in order to generate hashes that are near one another for images of the same class and far away for images of different classes. We evaulate the supervised and unsupervised variants of this model on the MNIST and CIFAR-10 datasets, as was done in the original paper, as well as a recent malaria diagnosis dataset from the NLM.