Portfolio
Below is a small selection of the plethora of scripts, projects, and little repos I wrote over the years. For a deeper dive into my work, please visit my GitHub page.
Ober Pipe Condition Classifier
A Convolutional Neural Network sequential model, trained from scratch on drainage survey footage. We are using Keras library for image classification based on condition of the drains and then export it as a Tensor Flow Lite model.
Ober Civils Toolkit
Civil Engineering Toolkit for AutoCAD: The Toolkit helps improve efficiency of civil engineering tasks within AutoCAD, focusing on the day-to-day needs of residential engineers. It streamlines a wide range of routine tasks and also offers automation for various non-residential but essential operations. It is widely adopted by numerous civil and residential-structural engineers in the field.
Polynomial Curve Fitting
Here’s a quick exercise in curve fitting to reverse-engineer an old CWI/SAAR plot from NERC (1975) Flood Studies Report (FSR) - Plot of catchment wetness index, CWI, against mean annual rainfall, SAAR. We aim to replace repetitive graph referencing with a single equation. We work out a polynomial expression that will get the CWI value for a given SAAR. All we are using is pandas, numpy, and matplotlib for the plots.
Chorochromatic Map Digitiser
This project uses Convolutional Neural Networks (CNNs) to digitise scanned historical chorochromatic maps, converting them into digital spatial data. The approach is applied to the Winter Rainfall Availability Potential (WRAP) map.
Quick multiplication by two with TensorFlow
We’re training a simple neural network, with one neuron at both the input and output, to double numbers in NumPy arrays, multiplying them by 2. It obviously learns to do this for any new numbers it encounters, without ever seeing the multiplication formula.
Get in touch using the form below if you’d like.