Module 3. From Observation to Evidence: How Data Becomes Meaning

Course 2: What is Data? Understanding the Building Blocks of Knowledge
Estimated Time: 25–30 minutes

🧭 Module Objectives

Explain how data move through stages from observation to interpretation.
Distinguish between "raw," "cleaned," and "modeled" data.
Describe the role of human judgment in every step of data creation.
Recognize how bias, context, and method shape what counts as evidence in the humanities.

From Seeing to Knowing

Every dataset begins as an observation of something in the world: a stone tool, a sentence, a song, a vote, a gesture. But data are not "found" like pebbles; they are made through selection and recording.

A simple chain of transformation can be summarized as:

Observation → Recording → Organization → Analysis → Interpretation

At each stage, something is added or lost. Humanists care about those changes because they carry ethical and cultural weight.

"Raw" Data Isn't Really Raw

The phrase "raw data" suggests something pure and objective. In reality, data are always cooked by decisions about what to observe and how to record it.

A scribe chooses which deeds to copy into a charter roll.
A musicologist decides which recordings count as "folk" or "protest."
An archaeologist decides where to dig, what observations should be recorded, and materials should be kept for further analysis.

Each choice filters reality through a lens of interest, training, and available tools.

As Johanna Drucker puts it, "data are capta" — things taken/captured, not things given.

Cleaning and Modeling: Making Data Usable

Once collected, data often need to be cleaned and modeled so they can be shared or analyzed.

Stage	Purpose	Typical Action	Humanities Example
Collection	Capture observations	Field notes, scans, recordings	Photographing inscriptions
Cleaning	Remove errors or inconsistencies	Correct dates, standardize names	Normalizing place names in letters
Modeling	Structure data for use	Assign categories, build relationships	Linking songs to albums and themes
Analysis	Extract patterns or insights	Count, compare, visualize	Tracking recurring motifs in lyrics
Interpretation	Tell a story or form an argument	Contextualize findings	Essay on how Welles revives protest folk

Even a spreadsheet of Jesse Welles' songs is a model: someone decided which columns matter—date, album, topic, tone—and which did not.

When Data Become Evidence

Data turn into evidence only when used to support a claim.

Field	Typical Claim	Data as Evidence
History	"Public sentiment shifted after 1918."	Letters, newspapers, polls.
Archaeology	"This building was used for rituals."	Artifacts, layout, organic remains.
Musicology	"Modern folk revivals draw on 1960s forms."	Chord progressions, lyrical themes, archival audio.

The same dataset can support different interpretations depending on method and framework, which is why the humanities emphasize transparency and context over mere replicability.

Bias and Perspective

No data are neutral. Choices of collection and categorization embed values and assumptions.

A map that centers Europe is a data bias.
A museum label that calls an object "primitive" is a semantic bias.
An algorithm that prefers trending content is a design bias.

Recognizing bias does not invalidate data: it reveals their human dimension. Humanists excel at this critical awareness.

Key Takeaways

Data are not neutral records but constructed artifacts.
Each stage—collection, cleaning, modeling—shapes possible meanings.
Data become evidence only through argument and interpretation.
Human judgment and bias are inevitable and must be acknowledged.

What is Data? (Data Intro Course)