I found a dataset of 13,000+ free campsites on USCampgrounds.info (major thanks to Tom Hillegass, the site owner) and decided to build an application that allows users to filter campsites and explore on a map.
Currently, the application consists of a PostgreSQL database (with a sqlalchemy ORM), a REST API using the python library FastAPI that allows for server-side filtering/sorting/pagination, and a React/TypeScript frontend with a filterable list view. The repo contains a Docker setup for running the database, API, and frontend locally.
I've completed some really cool work on adding a Places functionality to the application. It's currently not included in the frontend, but can be accessed directly in the API here. I downloaded all of the geographical names data for the US and Canada (which is an unbelievable amount of data) and filtered it down to major places of interest, such as cities, lakes, and mountains. I then added PostGIS to the database and updated the campsites data to include a new Point column with its coordinates. The Places API supports search by string or location as well as search by distance from another point. The Campsites API also supports searching by proximity to a given point. This will support a more intuitive search functionality for finding campsites as well as the ability to plot and cluster campsites on a map.
Developed by request in October 2021. A middle school science teacher needed a replacement for a Flash app she had used to simulate predator-prey relationships in ecosystems. She wanted a website that would have two dropdowns, one showing predators and another showing prey, and feedback about the choice.
I would say this project is most interesting for its management and validation of JSON stored locally, as well as upload and download capabilities. I also wanted to practice using Material UI v5 as I've only worked with v4. On the shortlist of feature-adds is a user interface for adding data (instead of forcing the user to work directly with JSON). I would like to add a simple backend to this eventually.
I received an M5Stack as a gift and wanted to find a nice project that would combine the Faces modules, button presses, using the SD card, and drawing graphics on the screen. I decided to implement a hangman game that would pull from a list of random words and show feedback on which letters had been correct and incorrect, as well as show a hangman which slowly fills up as the player guesses incorrect letters.
I used this as an opportunity to practice splitting a one-file monster into multiple, organized files. I'd worked with Arduino extensively in the HCI lab but had only ever put my programs into a single file (and for CRISPEE that file was very, very, very long). This seemed like a good chance to think about splitting the project into three logical chunks - the business logic, the graphics, and the SD card reader. Code exists for a version that uses the Merriam-Webster API to fetch the definition of the secret word, as a hint for the player, but this was not completed to the point of being ready for publication.