Built something? We create video reels & spotlights for GitHub projects.Promote your project →
>GitHubRepo
Promote your project
Home / Developers / theusual
Developer Profile

theusual

Discovered public repositories for theusual in the GitHub catalog.

↗

theusual/kaggle-seeclickfix-model

Prize-winning solution for Kaggle contest "SeeClickPredictFix". The purpose of the contest was to train a model (as scored by RMSLE) using supervised learning that will accurately predict the views, votes, and comments that an issue posted to the www.seeclickfix.com website will receive. This code developed by me uses a segment based ensemble to generate predictions for the 3 targets (views, votes, and comments). My teammate and I used this model in combination with his to create a winning model for the contest, defeating >500 other teams and winning a purse of $1,0000.

↗

theusual/kaggle-seeclickfix-ensemble

Prize winning solution to the SeeClickFix contest hosted on Kaggle, developed by teammates Bryan Gregory and Miroslaw Horbal. The purpose of the contest was to train a model (as scored by RMSLE) using supervised learning that will accurately predict the views, votes, and comments that an issue posted to the www.seeclickfix.com website will receive. My teammate and I used this ensemble code base to combine our top ranked individual models to create a prize-winning solution, defeating >500 other teams and winning a prize of $1,000.

↗

theusual/reverse_geocoding_nominatim

Quick hack to perform reverse geocoding on the Nominatim map database using longitude and latitude. Currently only pulls address information (street name, zip code, city, neighborhood) but could easily be updated to pull other info as well (country, country_code, state, and county). Also, currently switches automatically between the OSM hosted nominatim database (http://nominatim.openstreetmap.org/) and the MapQuest hosted nominatim database (http://open.mapquestapi.com/nominatim/v1/). Switching is performed when one times out or when a threshold is hit, to avoid overloading either of the servers.

↗

theusual/mobile-app-missionary-helper-cambodia

Mobile HTML5/JS based app for Iphone/Android/etc. Designed to assist missionary's in their travel's to Cambodia by providing sound clips in Khmer for various bible verses, words, and phrases. Also has local emergency contacts, maps, and other resources.

↗

theusual/kaggle-yelp-business-rating-prediction

Code for Kaggle Contest "RecSys2013: Yelp Business Rating Prediction" -- Predicting Business Ratings on Yelp.com Using Machine Learning. My final submission scored an RMSE of 1.2307 which earned a rank of #7 out of ~250 teams. This a collection of code that is meant to be run from python console, it is not a stand-alone program

↗

theusual/mobile-app-dont-break-the-chain-web-version

Don't Break The Chain -- A mobile app for helping users easily and enjoyably track progress towards goals. Useful for both making and breaking habits, the user can graphically mark each day that their daily goal is achieved, creating a "chain" of days representing success. The app is available across all mobile platforms and works both online and offline. All goal progress synchronizes properly across all platforms. This is the code for the WEB version of the project, hosted on a web server and accessible via a web browser.

↗

theusual/sql-scripts-macola-wisys-epicor

A backup repository of my SQL scripts used both inside and outside of the workplace including: -Helpful DBA scripts -Macola Progression scripts -Macola ES scripts -Wisys scripts -Brysys scripts -Epicor scripts

↗

theusual/kaggle-yelp-useful-vote-prediction

Code for "Yelp Recruiting Competition". The goal was to predict the useful votes that a review posted on Yelp.com will receive using a predictive model created using supervised machine learning. This was my first Kaggle contest and I entered late so it was drafted in a rush, only was able to submit one entry ont he last day. It scored an RMSE of 0.53384 which landed me 135th out of 350 teams. I was pretty happy wiht that given that it was a last minute submission and my first foray into machine learning.

FOR MAINTAINERS

Built something? Put it in front of millions of developers.

We make a short reel about your project and post it across YouTube, Instagram, Threads, and X. Send a link, we do the rest.

Promote your project