Puntofisso blogs here

15 things I have learned launching AI projects - Part 3

Tue, 07 Apr 2026 13:00:00 +0000

Systemic issues in AI adoption

In the previous issues of this multi-part blog post I shared some lessons in AI adoption that pertain to definition and framing of the opportunity. We’re now looking at the systemic, structural issues that can impede proper adoption. To note, I don’t go into technical details in this blog series (maybe one for later?). But if you want to learn more about the technical issues and opportunities that a technical adopter may face, head to Simon Willison’s blog. I’ve learned a lot of my tech skills in AI from him.

15 things I have learned launching AI projects - Part 2

Mon, 30 Mar 2026 09:02:00 +0000

What does AI adoption means in practice

Part 1 was a preamble based on framing AI in the context of building solutions. Here, we get a bit more practical detail.

15 things I have learned launching AI projects - Part 1

Mon, 30 Mar 2026 09:01:00 +0000

The context in which we build AI

This is a little preamble because it’s mostly about how to frame AI in a way that makes you… forget about AI.

15 things I have learned launching AI projects - Index

Mon, 30 Mar 2026 09:00:00 +0000

In 2020, I joined the NHS AI Lab. Part of the now defunct NHSX (a joint digital unit between NHS England and the Department for Health and Social Care), the AI Lab was a £250M initiative to develop and deploy safe, ethical AI across the UK healthcare system.

All I Want For Christmas Is Davos

Fri, 20 Feb 2026 08:14:36 +0000

I bumped into this nice visualization of Google Trends. It’s about web searches for “Davos”, which clearly exhibit a seasonal pattern consisten with the dates of the World Economic Forum.

Using a M5AtomS3R to display live bus arrival info

Sun, 11 Jan 2026 16:45:00 +0000

I’ve been meaning to play with microcontrollers for a long time. When I was at university, I was really jealous of my friends who had studied electronics in high school and were able to use these mysterious devices. The learning curve was really hard before ESP32 and Arduino made things much more standard and easy. The latter, especially, came with an IDE that, over time, has become the de facto standard not just for the Arduino family of devices, but also for a range of others, supported by installing external libraries.

The Bus Station That Didn’t Exist, and Other Data Epiphanies

Mon, 27 Oct 2025 12:32:13 +0000

This article was first published in Nightingale

“Data is multidisciplinary” is my mantra—it’s 2025, and I’ve now worked 20 years in every possible flavour of data—data visualization, open data advocacy, data pipelines in healthcare, data-driven national-scale services, AI innovation, and more. Whatever the application or project, my take on data literacy is the fundamental ability to challenge your own assumptions about the data you have or don’t, the appropriateness in using it, the ethics of your application, and ask yourself: is there a different way, perhaps? Here is a gallery of some of my most treasured eureka moments working with data.

Notes from csv,conf,v9

Mon, 15 Sep 2025 16:00:00 +0000

I’ve just come back from a splendid week in Bologna, where I helped run csv,conf,v9 for the first time.

Commutes within urban areas are the shortest

Sun, 14 Sep 2025 19:00:07 +0100

A couple of weeks ago I saw this interesting map of commute times in Germany made by Datawrapper’s co-CEO David Kokkelink, noting how commutes into major cities are the longest in Germany.

The Open Data Delusion

Fri, 06 Jun 2025 11:29:07 +0100

Photograph: The Helen Hamlyn Centre for Design

Open Data is about the “Data” as much as it is about the “Open.” Some stories from my experience as an Open Data activist and adviser illustrate it.

QGIS – An ATLAS of buildings by council

Thu, 23 Jan 2025 14:45:00 +0000

Another little QGIS step-by-step note to my future self.

Prerequisites: This time I was looking to create a visualization of London councils showing the buildings that come from Open Street Map, using the handy downloads from Geofabrik to get the buildings shapefile into QGIS. In order to add the “by council” element, I also downloaded an official shapefile of the boundaries of London councils (e.g. from here) and then joined the two layers using QGIS geoprocessing tools. I’ll leave this to you as exercise, but the final output is a layer with the buildings and a local authority unique ID.

Playing With Gpx Data from Strava

Sat, 11 Jan 2025 10:45:00 +0000

I’ve been wanting to try using gpx data for a while. You can record your runs with a variety of apps, and even edit those with a text editor, as gpx is a format that is relatively human-readable. What I realised, though, is that I’d been using Strava quite a lot when we were in lockdown in 2020, my hour of daily air being, often, a jog. Personal best after personal best (which didn’t take much, as I’d never really been a runner before) I got to know areas close to where I lived that were suitable for running, and also venturing for longer runs, including some over 10k.

Three things about data...

Fri, 06 Dec 2024 16:45:00 +0000

First published on LinkedIn

Quite a few interesting discussions at Think Data yesterday, both on and off stage.

Calculating the average face from a set of photos using OpenCV on Colab

Sat, 16 Dec 2023 16:40:00 +0000

TL;DR Here is the Colab notebook.

A few years back, the the UK Parliament released photo portraits of each Member of Parliament. So, I thought, it would be cool to do something data-driven with that image set. I had worked before with algorithms that allowed to find reference points on a face, using Terence Eden’s code that found the most similar painting to a face.

Making a map of the closest capital using QGIS

Fri, 18 Aug 2023 10:20:00 +0000

I saw this map on Facebook. It asks what capitals are closest to each Italian town, creating what looks like a continuous map to display the result. I like this kind of thing, so I set out to replicate it using QGIS, and making the process as replicable as possible.