3D-viz revisited, Swedish sentiment and bad AI

The J++ newsletter is back! From now on, the newsletter will be dispatched either in English, or bilingually in English and Swedish. We start this year by giving the dreaded 3d vizualisation another chance, with a few new tools, and a look at how The New York Times did their December story on smartphone data collection.

Your phone apps share your whereabouts for profit
Since Edward Snowden released classified information on global surveillance programs in 2013, the debate around privacy is continuously brought to the forefront of the public discussion.
In Your Apps Know Where You Where Last Night And They're Not Keeping It Secret, The New York Times received a database of more than a million phones in the New York area. And despite the data being anonymous, it was so detailed (home address, workplace, healthcare centre, relatives...) they could tie it to individuals. This database is an example of the data sold by apps companies for location-targeted advertising (a booming market with an estimated $21 million in sales in 2018). But when users share their location data, they aren't informed it can be sold, aggregated, stored and analysed:
"Some companies say they delete the location data after using it to serve ads, some use it for ads and pass it along to data aggregation companies, and others keep the information for years."

Analysis of location tracking companies
In connection to the investigation mentioned above, reporters of the New York Times tested 10 apps (iOS and Android versions) that collect precise location data (How The Times Analyzed Location Tracking Companies). The reporters used research softwares to identify the data sent over and the companies receiving the data to compare it with their often blurry privacy policies.

Inspiration

Data visualisation in 3D
Why look at dots on a map, when the 3D view can give us a better understanding of how people are distributed throughout the world?
The Pudding used data from the Global Human Settlement Layer to present population density based on a post by Alasdair Rae.
Source: https://pudding.cool/2018/12/3d-cities-story/

Election results using a cartogram
To display the results of the recent US House 2018 election, the New York Times published an article where they let users to choose between a map and a "cartogram". The latter choice gave rise to interesting feedback (see Michael Sandberg's data visualisation blog post).
Cartograms display distorted geographical information to allow for the display of another variable. In this case it show the number of seats for each party. And not only are the states apart while still reflecting the global shape of the country, it is also a very informative representation. From glancing at the cartogram (as opposed to the map) it is obvious that the Democrats (in blue) got the majority of seats.

Tools

SenSALDO - A Swedish sentiment lexicon
Tools and methods exists to analyse the sentiment within a text, but most are based of and optimised for the English language. SenSALDO is a sentiment lexicon for Swedish based on the full-scale Swedish semantic lexicon for language technology applications (SALDO) developed by research programs funded by the Swedish Research Council.
Source: https://spraakbanken.gu.se/eng/resource/sensaldo
Publication: http://www.lrec-conf.org/proceedings/lrec2018/pdf/857.pdf

Jupytext - Jupyter notebooks as markdown documents
Jupyter notebooks are great interactive tools for experimenting with data and code and sharing the result. Jupyter notebook depends on a underlying JSON based format. to be displayed and edited, well not anymore. Jupytext is a tool that enables you to write notebooks as scripts or Markdown in any text editor or as a .ipynb file that can be edited outside of Jupyter.
Source: https://towardsdatascience.com/introducing-jupytext-9234fdff6c57

Miscellaneous

Awful AI - Raising awareness to the misuses of AI
Artificial Intelligence (AI) is getting more widely used by companies and governmental agencies around the world despite evidence of biases. Awful AI aims to track the current misuses of AI in the hope of triggering a discussion around opposing biased AI tools.

J++ Stockholm | http://jplusplus.org/sv/

Want to change how you receive these emails?
You can update your preferences or unsubscribe from this list.