472: quantum of sollazzo

#472: quantum of sollazzo – 14 June 2022

The data newsletter by @puntofisso.

Hello, regular readers and welcome new ones :) This is Quantum of Sollazzo, the newsletter about all things data. I am Giuseppe Sollazzo, or @puntofisso. I’ve been sending this newsletter since 2012 to be a summary of all the articles with or about data that captured my attention over the previous week. The newsletter is and will always (well, for as long as I can keep going!) be free, but you’re welcome to become a friend via the links below.

Thanks to COVID (10 days positive… and counting), I had some time at home, which I spent, among other things, getting started on creating a basic cartogram or hexgrid map of the ~1500 NUTS3 areas in Eurostat. It’s a work-in-progress or, rather, work-that-I’m-not-so-sure-whether-to-progress :) If you think this might be useful, let me know, and I’ll keep finalising it and share it as open source.

I really enjoyed taking part in a panel about AI at the BBC. Well, except I had to miss the drinks part thanks to said COVID… however, the discussion, about responsible AI in public services was pretty good, with Myrna from Ocado and Bill from the BBC giving incredibly interesting perspectives to compare with my own. Apparently, a recording will be shared soon so stay… tuned. Yes, the feeling is pretty much Big Rottweiler Brother is watching.

Every week I include a six-question interview with an inspiring data person. This week, I speak with Giorgio Comai of OBCT/EDJNet. His data wrangling tutorials are among my favourites.

Speaking of EDJNet, in the second half of 2022 they will launch a series of free introductory courses to data journalism in collaboration with Dataninja (whose co-founder Donata Columbro was on our “Six Questions” in issue 448). The best part is that they are entirely co-designing the courses, and your opinion is highly valued – just fill in this survey. The courses will all be in English.

‘till next week,
Giuseppe @puntofisso

Six questions to...

Giorgio Comai

Giorgio is a data analyst and researcher at OBCT, crunching data for EDJNet.

What is your daily data work like and what tools do you use?

I think of all of us working with data as a big hunter-gatherer society. Sometimes our work resembles that of gatherers, i.e., we rely on open data or pre-existing datasets. In other occasions, we are more like hunters.

In my stereotyped image of such societies, there are folks who remain somewhere quiet in the shadows, building the tools that others use for hunting and gathering, processing the food they bring home, and making sure it’s of good quality. This is what I mostly do nowadays.

In brief, I spend my working days with the RStudio IDE in front of me.

Tell me about a data project that you're proud of...

Besides double-checking datasets, it’s really the tools and tutorials that I feel are my most useful contribution in the data space, even if - due to time constraints - they are never quite as complete as I wish they were. For example, I am quite happy with how an R package I made to streamline the retrieval of data from Wikidata has been maturing. But I am also really fond of woefully underdeveloped projects, such as when I began working on a modern implementation of the Risk Characterisation Theatre: the package I started developing at the time is barely functional, but I still look at it as a meaningful contribution to the ever-important debate on how to most effectively communicate risk and uncertainty.

...and a data project that someone else did and you're jealous of.

I admire a lot R packages with elegant APIs, such as gganimate or rayshader. At first, animations and 3D renderings such as the ones they produce may just look gimmicky, but the APIs themselves push you to think more consciously about space, time, and layers in your data, which is good. Data-based animations still struggle to find their place in “serious” data visualisation, but I feel their time will come, perhaps as a partial replacement for limited interactivity or as an intermediate step towards data physicalization.

If I say "dataset", you think of...

… its authors. Who collected the data, and why, and… how did this dataset reach me? Did I produce this dataset myself from a combination of sources? Was it hunting or gathering? No matter the response to these questions, even if it’s nicely shaped (particularly if it’s nicely shaped) , even if I created it myself (particularly if I created it myself!), I look at it with suspicion.

Give someone new to data a tip or lesson you wish you'd learned earlier.

Testing. As soon as you decide to use a dataset, throw some quality tests at it to check for consistency. When you process data, introduce automated testing at different steps of the data-processing pipeline (this can be done even if you are working with spreadsheets). You end up asking yourself better questions about the dataset, you are often forced to clarify what each data point actually represents, and you substantially reduce the risk of getting things wrong. But of course, you should also be suspicious of your own tests, don’t rely on them too much.

Data is or data are...

Here’s an important truth about data, hidden in plain sight in this very question you keep asking: different people look at the same question, even starting with the same data, and, in all honesty, to the best of their knowledge and capabilities, will reach a different answer.

In situations when there are different legitimate ways to look at the same question, as data analysts, we should mainly take one of two roads. Either we clarify and qualify the question, until we feel we would all reach the same answer. Or we embrace and cherish complexity, outline and present our approach, and hope others will enrich our understanding with different approaches, different answers.

Topical

The war in Ukraine is fuelling a global food crisis.

The situation, as described brilliantly in charts by Reuters, is rather grim, to put it mildly. And, to compound the effect, the food crisis is linked to the energy crisis.

Un puente que resume 100 días de guerra: “Esto va más allá de la estupidez, es sabotaje

Interesting from El Confidential, and automatically translated into English, “A bridge that summarizes 100 days of war”, with a great animated visualization attached.

Midterm elections 2022: The issues that matter to Americans

“As the 2022 midterms approach, see which issues people in your congressional district care about.“
I’m honestly salivating at the beauty and richness of this data.

Contracting Emissions

Spend Networks has released an interactive data analysis of carbon emitted by the public sector’s supply chain the carbon emitted by the public sector’s supply chain, calculated using open procurement data.

Support For Gun Control Will Likely Rise After Uvalde. But History Suggests It Will Fade.

FiveThirtyEight looks at polling data about support for gun control, suggesting that, while it is unsurprising that Congress will do nothing, public opinion will always soften on the issue as time goes by.

Tools & Tutorials

What’s the most common crop in the US?

Six Questions graduate Erin Davis looks at US crops, with her usual approach based on a collection of great maps and R code.

Improving society with open data

Microsoft has published this curated data portal which purports to help people to “explore open datasets and learn how other researchers have used them to solve societal challenges.“

Dataviz, Data Analysis, & Interactive

System

“System is a free, open, and living public resource, being built by and for the world, that aims to explain how anything is related to everything else.“
An easy challenge, no?

Timely Advice – How Long Does Dataviz Take?

A dataviz and data analysis on… dataviz.

Wie weit komme ich mit dem 9-Euro-Ticket?

You might have heard that the German national railway company has issued get-me-anywhere 9-Euro tickets for this summer. In this interactive article, the Berliner Morgenpost looks at where those tickets can get you, in German and automatically translated into English

AI

Ethical AI isn’t just how you build it, it’s how you use it

“What good is removing biases from a robot that turns grandmothers into smoothies?“
Just as I was on a panel on responsible AI last week, this appeared in my inbox.

quantum of sollazzo is supported by ProofRed’s excellent proofreading. If you need high-quality copy editing or proofreading, head to http://proofred.co.uk. Oh, they also make really good explainer videos.

Sponsors* casperdcl and iterative.ai Jeff Wilson Fay Simcock Naomi Penfold

[*] this is for all $5+/months Github sponsors. If you are one of those and don’t appear here, please e-mail me