Portfolio

---

Recent Publications

• Robocop, Skynet, and Talking Toasters. Automation, AI, and Augmentation in policing (Policing Insight)

• How to Train your AI (TechUK)

• Analytic Technology Industry Roundtable Study: Analytics and Use Cases (Mitre)

---

Knowledge Graph Building from NLP using Spacy

A simple evaluation looking how off-the-shelf NLP models and tools can be used to produce a Knowledge Graph. Large amounts of data are in an ‘unstructured’ or semi-structured format and to run conventional Link Analysis analytics requires the data to be processed into a structured format. To achieve this transformation I have defined the following processing pipeline

Named Entity Extraction -> Co-reference Resolution -> Semantic type assignment -> Relationship Extraction -> Knowledge Graph

---

Data Cube Visualisation

Experiment with 3 dimensions, the following Jupyter example takes the gps output from my Garmin Vivosport and an Ordnance Survey basemap plotting latitude and longitude on the x & y axes and time on the z axis. I compare this to the standard Garmin Connect output which, using 2 dimensions, highlights pace a a colour map [green to red]

3D Visualisation	2D Visualisation