Joshua O’Brien

This site is a dynamic Django-powered portfolio where all projects, images, and skills are managed through the admin panel.

V1, as mentioned below, was originally built in C#, Blazor, and Azure, then rebuilt in Python and Django to align with my core tech stack. The site is hosted on AWS Lightsail and served through Nginx and Gunicorn.

I built a fully automated property market reporting platform powered by a serverless data pipeline.

AWS Lambda processes clean and load Land Registry and postcode datasets into Snowflake. When a user enters a postcode, the website triggers an API-driven Lambda that queries Snowflake, runs Python analytics, and generates a tailored PDF report.

To improve performance, monthly reports are cached in S3, enabling instant retrieval for repeat requests and reducing compute overhead.

Leading the architecture and automation of the data infrastructure supporting a €3.6bn real estate finance portfolio - some projects discussed on this site.

I design and maintain Python-based ETL pipelines, custom API integrations, and event-driven workflows on AWS to deliver accurate, automated reporting and analytics. These systems streamline operational processes, reduce manual workload, and provide stakeholders with faster, data-driven insights.

My first published iOS app was built for a Design and Construction firm who needed a mobile extension of their project management portal.

The project began in August 2023 and involved working closely with the company’s web developer to design and implement custom APIs, enabling the app to read from, write to, and update their existing database in real time.

The final product was officially named “Bulldog Hub” by the client.

As a requirement to register as an Apple Developer, I need my own website (I always wanted one too!). So, I created one using Squarespace... for the low price of £19.20 pcm (£230.40 pa). This price was eating away at my profits and was returning Zero value.

I delved into web development and static web hosting on Azure. Starting with a Microsoft Blazor template and drawing inspiration from my original Squarespace design and content, three days later, we are here!

A fully functioning website that has £0.00 pcm cost as the site qualifies for the Free Tier on Microsoft Azure!

Using open source crime data provided by the FBI, I was able to make an interactive visualisation of that data using PowerBi.

This involved the initial configuration of the Excel based data tables into a structure that was easy to use with PowerBi. Once this was done, I then linked with PowerBI and used both built in and third party visualisation tools to create what you can see in the image. (More pages included in the final report)

With this interactive visualisation, my client (a none paying friend) was able to quickly and efficiently communicate trends and information, while slicing the data on the go.

Newmark UK (Gerald Eve pre Newmark Acquisition)

Context: A single excel model is used to value £100m to more than £1bn of real estate using the DRC method. We do the math, and build the report all within the model. The report is generated by duplicating a template sheet for every asset, then printing the print range on each sheet.

Problem: The VBA based macro takes 2 - 3 hours to duplicate sheets based on the template sheet (100s of sheets per Excel, taking 2 mins per duplication)

Developed Solution: Python based .exe that allows you to select the excel model via file explorer, and run the python code, of which used OpenPyXl to duplicate the sheets without opening the Excel file - a rapid solution that takes a matter of 3 minutes for the largest model at the time.

A personal project that arose out of my job. We purchase PDF documents from the Land Reg and have no app to store all of these.

This solution uses an AWS SQL database to organise and store (as BLOBs) the Land Reg documents we own. I used Python to build the programme, utilising custom functions as well as PYQT5 for the interface and Requests for communicating SQL commands to the AWS Database.

Cool feature: Auto reading PDF contents to autofill required fields (and validate documents), which allows for automatic Bulk uploading of documents. (Uses Python PDF Reader Library).

It can scrape the Land Registry interactive map (using selenium) to check if an asset shares a Freehold / Leasehold title, which we can then link in our database - however this breaks site ToEs, but is very useful…

10 years of monthly PDF based data in a tabular layout, converted into an interactive data visualisation.

I did this by using a Python PDF reader library to loop through all 10 years of PDFs, covert the specific tables into .csv files, extract the desired data from each .csv and store in an Excel based database. I then connected this database to PowerBi and utilised the visualisation tools to produce the what you can see (page 1 of 2).

I also created a Python tool which can be used to automatically add new monthly data to the database which is produced every month.

Placeholder

My first fleshed out iOS app for carrying out Red Book / International Valuation Standard calculations. Developed in 2020/21 based on the Real Estate BSc - Year Three Valuation module. Features include:

- Saving Valuation Calculations using Core Data
- Programming of Valuation Algorithms (E.g. DCF, Hardcore Layer, etc.)
- Custom Side Menu

Not the perfect app and it certainly has its flaws! However, all lessons learned for future developments.

Placeholder

Placeholder

Placeholder