# Case Studies

Two worked research scenarios showing how the Swiss-AL tools combine in practice.
Each follows a researcher from an initial question through to an analysis, naming
the tool used at every step.

## Climate Change

**Tools used:** [Semantic Space](explore-corpus.md#semantic-space) ·
[Context of Words](search-corpus.md#context-of-words) ·
[Distribution of Words](search-corpus.md#distribution-of-words)

Susanne, an environmental scientist, is researching climate change — a widely
covered topic — but she has no experience collecting large text datasets or
crawling news websites. She needs a way to navigate the Swiss media landscape
quickly and identify relevant terms and trends for further analysis.

She wants to know which terms are most commonly used when reporting on climate
change in Switzerland, and how different media outlets write about the topic.

### Step 1 — Exploring vocabulary with Semantic Space

To discover terms related to climate change, Susanne uses **Semantic Space**. She
searches for *Klimawandel* ('climate change') in the
[German journalistic corpus](corpora-journalistic.md), which covers high-reach
German-language media published since 2010.

```{figure} _static/scenario_klimawandel_emb.png
:alt: Semantic Space results for Klimawandel, showing nearest neighbours
:width: 700px

Nearest neighbours of *Klimawandel* in the German journalistic corpus.
```

The tool suggests near-synonyms — *Klimaerwärmung* ('climate warming'),
*Klimaveränderung* ('climate change'), *Erderwärmung* ('global warming') — as well
as contextually related terms such as *Klimakrise* ('climate crisis'),
*Klimakatastrophe* ('climate catastrophe'), and *Artensterben* ('species
extinction'). This gives Susanne a richer vocabulary to work with.

### Step 2 — Investigating usage with Context of Words

Next she examines how these terms appear in articles. She opens
[Advanced Search](advanced-search.md) and enters a CQP query:

```
[lemma="Klimawandel|Klimaerwärmung|Klimaveränderung"]
```

This finds all word forms of the three lemmas. She can then explore the **Context
of Words** view, showing text snippets around each hit, and **Distribution in
Documents**, showing where in each document the terms occur. Every example links
back to the original article in the SMD Swiss Media Database.

```{figure} _static/klimawandel_klimaerwaermung_klimaveraenderung_kwic_view.png
:alt: KWIC view showing the three climate terms in context
:width: 700px

Context of Words view for the three climate terms.
```

```{figure} _static/klimawandel_klimaerwaermung_klimaveraenderung_text_view.png
:alt: Document view showing where the terms fall within each document
:width: 700px

Distribution in Documents for the same query.
```

### Step 3 — Comparing outlets with Distribution of Words

To see how frequently these terms appear across outlets, Susanne uses
**Distribution of Words**. This lets her compare relative frequencies by source
and identify which outlets focus most on the vocabulary she selected. With a
clearer overview, she creates subcorpora for selected media — *SWII*, *WOZ*, and
*WEWO* — for deeper comparative analysis.

```{figure} _static/klimawandel_klimaerwaermung_klimaveraenderung_distribution.png
:alt: Bar chart comparing term frequency across media sources
:width: 700px

Distribution of the climate terms by source.
```

## No Billag

**Tools used:** [Context of Words](search-corpus.md#context-of-words) ·
[Distribution in Documents](search-corpus.md#distribution-in-documents)

A media scholar wants to understand how the current debate about the
*Halbierungsinitiative* ('Halving Initiative') works. To do so, he looks back at
how the *No Billag* referendum was reported at the time. His hypothesis is that
appeals to democracy, neutrality, and cost savings were already in use then as
argumentative strategies.

His question: how do democracy, neutrality, and costs appear in connection with
*No Billag*?

### Step 1 — Getting an overview

He searches the German journalistic corpus for *No Billag* and *No-Billag* in
**Context of Words**, which gives him source, title, date, and surrounding text.
It is immediately clear that *No Billag* is by no means always the main topic of
an article that mentions it.

```{figure} _static/kwic_no_billag.png
:alt: KWIC view of results for No Billag
:width: 700px

Context of Words results for *No Billag*.
```

### Step 2 — Separating mentions from coverage

To find the articles that actually deal with *No Billag* rather than merely
mentioning it, he switches to **Distribution in Documents**. The contrast is
visible in the data: in No. 62, "Steuerreform: Es soll schnell gehen und ohne
Referendum" ('Tax reform: it should be quick and without a referendum'), the term
appears once in passing. In No. 3, "Wendepunkt in der Geschichte der SRG"
('Turning point in SRG's history'), it is the subject of the article.

### Step 3 — Reading the relevant articles

He then selects the articles relevant to his question — No. 3, 61, 36, 76 — and
follows the source URL to read them in full on
[swissdox.ch](https://swissdox.ch), which requires separate access.

```{figure} _static/swissdox_no_billag.png
:alt: A No Billag article displayed on swissdox.ch
:width: 700px

Following a result through to the full article on swissdox.ch.
```