Questions To Ask Your Data if You Are a Humanist (or, Just a Human)
This is the simplified version of a spec in the W3C style available in html and in Bikeshed mark-up. Everything is available on Github
During the academic year 2015 to 2016, the six Praxis Fellows developed ClockWork, a series of sonifications exploring the cost of consumer goods and wages. We chose to use questions as a way to record our inquiries and decision making process as we moved from an abstract, time-related concept to a completed project. Instead of a traditional and specific explanation of exactly what we asked and answered as a group to create ClockWork, this documents our journey, foregrounding the questions we faced, which are edited to be broadly applicable to any group working with and on humanities, data, and the digital world.
The editors formed a documentation team to produce this document. The role of the documentation team is to note the inquiries and decision making processes of the Praxis group as we move from concept to conclusion. This information aims to make our trajectory and final project transparent by highlighting how and why we reached our deliverables. It allows others to model their projects after this one. It serves as a teaching tool and offers a “behind the scenes” look at the nuts and bolts of creating a digital humanities project.
This standard draws on the style and technique of [[security-privacy-questionnaire]] and [[SocialJusticeandDH]].
Representations of Data
This section asks a series of questions about the representation of data. It was developed as questions regarding sonification—representation by sound—of data, so sees all representation through this lens first, but has been generalized to suggest humanistic questions about the representation of data in general.
The process we found useful in doing the work was to make a very quick draft and present the rough version to the group to guide each revision. The questions are divided into planning and assessment. The planning are questions we found useful to ask before creating a draft the assessment questions were useful to ask of each draft to guide the next version.
How will the data be extracted, organized, and edited to make it work within the representation system? Can the process be automated? Can you use off-the-shelf tools?
If the value fluctuates, how does the sonification/representation show those different values? What other sonic/aesthetic considerations will arise?
Should this produce something pretty or aesthetically appealing? For example: do we want commodity prices and a musical scale, or commodity price as stable sound (do) up a little (re)- then note (do) can be combined…? Are we making music?
How do the different changing values relate to each other in the sonification/representation? Which are linked in sound and which are not linked? Can we experiment with different linkages? For example: do we care more about how different incomes relate to commodities or how different incomes relate to each other? Can we try them different ways?
How easy is it for you to manipulate which variables?
How did you decide the timbre? Color? Pitch?
What do we like and not like about this sonification so far? Topics discussed include pitch/frequency scale, accessibly and speaker range, overlapping sounds, midi vs. frequency, need for preamble/explanation, pretty/palatable vs. harshness, what sound lends itself to information dissemination?
How would this sonification be seen within academic electronic music conferences? Or if a visualization to electronic art audiences?
How long did it take you to make each one?
How much of this can you reuse to make something slightly different?
How many versions do we want to present to make our points?
If sonic, can we make a visual version for accessibility? If visual, can we make a sonic version for accessibility?
What does this particular sonification/representation of data argue? For example: is the take away from the data that poor people have a harder time buying milk? How do we make sure that is not the (only) takeaway?
What is the range of time and activities that our sonifications/representations will represent?
How are our sonification/representation examples connected? What is the deliberate choice? For example: do we always have min wage as a baseline, then we add other variables?
Issues in Picking a data set for sonification
What can we represent using sound and how? What are we showing/highlighting/saying with our sonic representations?
Are we being journalistic? Editorializing? Making a point, or proving a thesis? Or, are we allowing a set of data to show us connections and information we did not know to look for?
Do we want granular blips or static pitches that change in amplitude? Harmonies? Mimetic sounds?
What lends itself to sonification in terms of scale? Do years, or decades, become minutes? How do we decide this and what is the significance or these decisions?
What is reliable data? Government? Corporate? Regarding our interest in prisons and prisoners: how do we acquire the data? What are the ethical implications of left out, non reported or under represented populations? How do we know what/whom we are missing?
In terms of workflow to create a portal and online home for the sonifications, what are the groups we need to delineate among us, and who is in charge of them? Group member(s) need to gather the data for sonification, sonify, document, portal/website building, curriculum building: who does what, and why? What are the areas of intersection between groups?
Team Organization and Planning
If we are doing sonification in March, when do can we create a website/portal to display, explain, etc. the sonification? Do we want a visual element with it?
How do we move ahead with the other tasks, e.g. how does curriculum move ahead without the sonification? Can teams work in isolation? What exactly do we want to do with this thing if we had the perfect tool? What does a successful deliverable look like?
How do we make accessible our replication process for sonifying data? Can we teach this? Is the bar of entry too high if we are using SuperCollider/coding languages? How do we make sure it is reproducible in Garage Band, Audacity, etc.?
Creating A Curriculum
This section outlines a set of questions to ask when creating a curriculum, teaching tool, or method to convey how one might replicate any portion of our sonification project. The general questions below apply to all humanists creating a curriculum and the more specific, nested questions reveal situations and issues we worked through in our own project.
- Who are we teaching? Who are the anticipated, desired, and possible audiences? How does this shape our curriculum building? Can we begin creating a curriculum without knowing who the user is?
- Are we interested in speaking to the digital humanities community with our curriculum?
- What tools and skills do we expect curriculum users to have? How do these tools and skills relate to age, gender, race, privilege, access to education and technology, etc.?
- Does our anticipated audience shape our design and functionality?
- What resources do we need to include to make our curriculum accessible for all anticipated users?
- What resources do we expect users to have?
- Can we address systematic biases through our curriculum design? Does linking to outside resources reinforce or combat, for instance, gender bias in coding? How do we create an environment for everyone to feel empowered and comfortable coding?
- What are we trying to teach, and why?
- Broadly, what are the primary spheres of our project and our curriculum? Which aspects of our project are we explaining, teaching, and contextualizing?
- What is implicitly prioritized or left out in our curriculum. What are the implications of these omissions?
How specific should our instructions be? For instance, do we make the SuperCollider sonification tutorial so that that people can input their own data? Or do we simply show them how we created our project?
- How do we create a curriculum that is interesting and alluring? Will anybody even look at it?
- Will people want to learn how to recreate projects like this one? Are we providing new ways of representing and understanding data?
- Where is the humanist element in our curriculum? How do we create a curriculum that marries technical tutorials and the humanities?
- How do we ensure that the humanities-based conversations regarding time, ethics, and representation that undergird our project are embedded in our curriculum?
- How is this project situated in digital humanities, as opposed to social science?
- How to we overcome/utilize the “spirit of critique” embedded in humanities scholarship?
Preparing A Portal
This section outlines a set of questions to ask when designing a web-based portal for a project. The general questions are broadly applicable to websites and digital projects. The nested questions offer a view into our conversations and debates.
- What questions are we asking or answering with this portal?
- How will the information and the way it is presented transform a conversation?
- Is it fruitful to design by targeting an imagined audience, like the digital humanities community? Is it fruitful to work instead on representations of data in time and space by imaging a user story instead of a user?
- What functionality do we want/need from our portal?
- What are the minimum features we need to effectively present our information?
- What is our reason for designing a web portal? Could we use Twitter, Tumblr, or another extant platform to display our information?
- Will our portal have artificial constrains, like no scrolling, no navigation, or a frustrating user experience? Why? Is there a benefit to these, or to a private, site specific web element? Will these inhibit access? Is this purposeful inhibition provocative in a useful way?
- How do we make the purpose of our design choices clear to users?
- What is our content and how do we label it? Who is in charge of what content and where is the overlap?
- Do we offer our curriculum, sonification, and documentation data via a traditional menu? Do we incorporate curriculum information throughout the site? Is this information embedded or does it open another window?
- How do we make our portal accessible?
- Should we create a tactile, 3D-printed version or visualizations to accompany our sonifications? Does built in frustrating user experience make the portal unfeasible for too many people?
- Are there fonts, colors, etc. that are poor design choices for accessibility?
- What is the design process and workflow like?
- Should we look through websites we like or don’t like? Should we use drawings? What guidelines do we need for usability and functionality. Would creating a wire-frame prototype help?
- What do we want the portal to look and feel like? Where and how do we share our information?
- What types of websites are there? Are long-form websites or game-ifications feasible with our current skills?
- Can we put the code on GitHub and into Jekyll so it can be group annotated?
Documenting a project makes your work reusable but also holds you accountable to understanding your assumptions and challenging unhelpful approaches. We vision a continuous process of documentation where one or multiple members collect the materials of discussion and decisions. This section helps ask questions to the group about the form that the documentation itself will take.
What is the process of getting documentation accepted by standardizing body that might be interested in this project? What is the appeal of having a standard? What is the process to submit the standard and how long does it take?
What do we mean when we talk about standards in documentation? Should the process be made standard at all?
What reaches the audience we want to reach? For example: how do we bridge mainstream web and DH communities?
What questions are people asking when doing their daily work that we’re thinking about?
Is the form we’re considering intimidating?
What exactly does access mean? What resources does someone need?
Do we want answers to the questions we are proposing that are essentially references, a bibliography? How does what we’re proposing improve on a simple bibliography?
How do we structure the documentation? For example: organize issues in the order of when the came up? Or, structure them into categories? (portal? data collection? sonification?)
Where is the overlap in documentation and curriculum regarding resources?
How do we implement our documentation page into our web portal? Do we want the whole web portal to be in the same format as the documentation page? Do we want a link, an embedded page?
Data Set Details
This section attempts to create a set of questions to ask when considering using a particular collection of data for a project. Not all the questions need answers, but each will be considered and any lack of knowledge about some aspect should be considered when assessing a data set.
- Where does the data come from?
- What do we understand to be reliable data? How are we acquiring this data and how was this data initially acquired?
- What are we doing when we use government data? How do we account for underrepresented and unreported information? How do we know what/whom we are missing?
- How transparent is the data? Are there undefined terms and unclear words? For instance, concerning BLS data, what is the difference between size A and B cities? What is meant by a “US Southern Average City?” Why are professions in which workers make over $200,000 per year not listed?
- What are ethics of data, data collection, and data usage as it pertains to our project?
How does the mechanism of access to the data work? Can it be queried, downloaded, processed?
- Who else can access the data? What special skills would they need? Institutional alliances? Legal or cultural capital?
- What laws, rules, or taboos apply to using the data?
- Is it public domain? Proprietary?
- Is it possible that someone may claim that you agreed to their license or that they own your results if you use the data?
- What cultures would regard using the data? Does some community have a claim on the data? What is the nature of the claim? (ethnic, spiritual, etc. etc.) What sort of voice does that community have in using the data? Have you informed them? Should you?
- How is the use, storage, and accessibility of this data being documented?
- Is your practice standard within a certain community?
- Where are you storing the data? Are you making it available in the form you acquired it? Or some other form? How long will it be available?
- How are you processing the data?
- Are you cleaning it? Changing it? Restructuring it? Changing the format? Extracting from it?
- Is the way you’re processing the data anticipated by the people who collected it? Does it harmonize with their assumptions about how the data will be used? Alter it? Does that change what the data seems to say?
- Are you documenting how you are processing the data? Does it reflect the standard approach of some community? Do they document their process and do you differ? Would someone finding the data know that?
- What tools are you using to process the data?
- Who can use those tools?
- What effect do those tools have on the people who collected, provided, stored, or are the subjects of the data?
- How is the data represented?
- Is the representation journalistic, editorializing, making a point, or proving a thesis?
- How does the representation alter what is prominent in the data? Does it bring attention to something that was unclear in the earlier forms? What does it bring attention to? How would the subjects, collectors, processors, licensers, cultural custodians, scholars, or other people think about that use?
- How will the data representation be put forth? Do you plan to inform people of your representation? Who? How? When? Where?
- What is the imagined impact of your data representation on people, organizations, etc. once it is public?
- What tools will be uses to publicize this data representation? Are these tools free? If not, what does the money support? Who has access to these tools?
- How accessible is the data representation?
- Can your project be seen as purely aesthetic by someone who doesn’t understand what you’re doing?
- What explanation or contextualization accompanies your data representation?