Home
Welcome 🍻🔗
I'm Gerard Bentley, a software engineer interested in the web, machine learning, and how both impact us. This is my blog and technical notes; feel free to poke around.
Currently working at Sensible Weather with a focus on Golang, Python, and Postgres webservices.
Most excited about time series data for environmental and social impact.
This site is generated from Markdown files with MkDocs along with MkDocs Material.
Some Demos🔗
Here are some side projects I worked on in the recent year or so.
Darts API Playground🔗
Explore the Datasets, Metrics, and Models of the Darts Time Series library.
See: Github Repo
URL Scanner🔗
Using AWS Rekognition + Streamlit to provide interactive OCR URL Scanner / Text Extraction on real world images.
See: Github Repo
🥞 WSGI Stack vs Streamlit🔗
Comparing an interactive web app built with bottle
+ htmx
to the same idea built with streamlit
.
In folder wsgi_comparison
🎥 Watch: Youtube Breakdown ✍🏻 Read: Blog Post
Left: ~50 lines of Python and HTML
Right: ~15 lines of Python
🎸 Guitar Tuner🔗
Simple guitar tuner powered by streamlit-webrtc
Streamlit Full Stack 3 Ways🔗
Demo of Full Stack Streamlit Concept. Deployed with 3 increasingly complicated backends.
See: Github Repo
Littlest🔗
Postgres Version🔗
Go Backend Version🔗
Fidelity / Personal Stock Account Dashboard🔗
Upload a CSV export from Fidelity investment account(s) and visualize profits and losses from select tickers and accounts.
See: Github Repo
💾 Pipreqs API + Streamlit Frontend🔗
FastAPI backend hosted on Heroku to parse repos and use pipreqs
to spit out a minimal requirements.txt
!
See: Github Repo
🚰 LightGBM Water Pump Predictions🔗
Data Science project based on DrivenData "Pump It Up" competition. Includes Data Exploration, Feature Engineering, and training and predicting functionality of water pumps with LightGBM
See: Github Repo
💰 Personal Spending Dashboard🔗
Upload a CSV or excel with at least a date column and spending amount column to analyze maximum and average spending over different time periods.
Peak Weather: NH 4,000 Footers🔗
Use async http request library httpx
to make 48 api calls roughly simultaneously in one Python process.
Feed a dashboard of weather for all 4,000 foot mountains in New Hampshire.
See: Github Repo
🐼 Pandas Power🔗
Demoing useful functionalities of Pandas library in a web app.
Currently:
read_html
: Parse dataframes from html (url, upload, or raw copy+paste)
✍🏻 Text Recognition Dataset Generator App🔗
Putting a frontend on TRDG CLI tool. Primary goal: creating classic videogame text screenshots with known ground truth labels
🐙 Github Lines of Code Analyzer🔗
Shallow clone a repo then use unix + pandas tools to count how many lines of each file type are present
streamlit run github_code_analyze.py
AWS Textract Document Text Scan🔗
Using AWS Textract + S3 + Streamlit to provide interactive OCR Web App.
See: Github Repo
Created: June 7, 2023