A Field Guide to Algorithms

What are algorithms? Who and what do they involve? What do they do? What is at stake with them? How can we account for them? How can we respond to them?

Following on from the Field Guide to “Fake News”, A Field Guide To Algorithms aims to gather and curate different starting points, recipes, approaches, experiments in participation and activities for collective inquiry into algorithms and the collectives, cultures, infrastructures, imaginaries and practices associated with them.

See also:

Digital Methods Recipes

How can we share different ways of doing things with digital data, methods and infrastructures? How can text, images, video, GIFs and other materials be used to provide accounts of digital methods, cultivate sensibilities towards interpretive work, surface tacit knowledge and encourage reflection on decisions, tools, devices, assumptions and materials?

This is a project to gather, exchange and share digital methods recipes and “how-tos” for research, teaching and collaborations from across the Public Data Lab.

You can see a preview of some of these here: http://recipes.publicdatalab.org/

A set of recipes developed as a collaboration between digital methods researchers at the Public Data Lab and digital journalists at First Draft can be found at: https://firstdraftnews.org/long-form-article/digitalrecipes/

Save Our Air

Supported by OrganiCity and developed by the Public Data Lab, SaveOurAir is an exploration of the social and political aspects of “smart cities”. Its specificity is the effort to use digital data to stir (rather than settle) urban debate and to nurture (rather than purify) their multiple attachments.

Focussing on air quality, SaveOurAir explored three ways to make urban data more “local” and “politically relevant” and developed three experiments in data activation.

MiniVAN

Networks are increasingly popular in the social sciences and in the humanities as interfaces for exploratory data analysis. “Visual Networks Analysis” (or VNA) allows scholars to analyse large relational datasets without having to deal with the full complexity of graph mathematics.

Current VNA tools, however, are either too complicated or unable to handle large datasets. Hence MiniVAN, a VNA tool, accessible to scholars with little knowledge of mathematics or coding and yet capable to scale up to graphs of several thousands of nodes.

MiniVAN draws on previous open source (gephi.orgsigmajs.orggraphology.github.io) and research projects (climaps.eu and fakenews.publicdatalab.org).

MiniVAN is a project of the Public Data Lab with support from Sage Ocean Concept Grant.

You can read more about the project in this interview with Sage.

About

MiniVAN is a tool for the exploration of small and large networks, addressed to social scientists and everyone who is interested in graph analysis but has difficulties with graph mathematics. The challenge is to make sophisticated analytic techniques (clustering, spatializing, ranking, filtering, etc) available in a visual and user-friendly environment.

MiniVAN has several advantages over the existing software, but it’s key strength is a long-standing reflection on the conceptual basis of Visual Network Analysis. MiniVAN will not only offer a collection of analytic functions, but also guide it users aligning different visualisations and metrics in a ordered inquiry.

The development of MiniVAN will be uphold by a series of discussions with potential alpha-users, through the support of the Public Data Lab a research network on digital data and public interventions. The PLD gathers scholars coming from many leading digital methods centres from across Europe (both from the social and the information sciences). It also entertains a network of contacts with practitioners, journalists, civil society groups, designers, developers and public institutions, which will be mobilised to specify the different needs that MiniVAN will try to respond to.

MiniVAN is composed of two distinct open source web applications:

  1. The Analyser
    • Receives and parses a graph uploaded by the user;
    • Guides the user through the choice of the dimensions and the methods of the analysis;
    • Accompanies the user in the definition of the visual properties of the graph representation;
    • Produces a file bundle to be processed by the second application. (The file bundle produced by the Analyser can be hosted by any web server or by a third-party solutions – such as Github’s gists)
  2. The Visualiser
    • Reads the data contained in the file bundle produced by the Analyser;
    • Displays multiple visualizations of the network;
    • Provides selected metrics and information about categories and relational structures;
    • Allows the proposed views to be further filtered/tweaked/explored etc. by the readers;
    • Generate a code snippet that can be embedded in any website.

A few readings on Visual Network Analysis:

A Field Guide to “Fake News”

A Field Guide to “Fake News” and Other Information Disorders explores the use of digital methods to study false viral news, political memes, trolling practices and their social life online.

It responds to an increasing demand for understanding the interplay between digital platforms, misleading information, propaganda and viral content practices, and their influence on politics and public life in democratic societies.

It is a project of the Public Data Lab with support from First Draft.

The guide is freely available via the link below.

It is released under a Creative Commons Attribution license to encourage readers to freely copy, translate, redistribute and reuse the book. All the assets necessary to translate and publish the guide in other languages are available on the Public Data Lab’s GitHub page.

You can also find it on Zenodo here (DOI: 10.5281/zenodo.1136271).