Resources by Maximilian

Open Political Statements Germany 2021 General Election (2021 Bundestagswahl)

Open Political Statements Germany 2021 General Election (2021 Bundestagswahl)

Maximilian
Open Dataset
Creators: Maximilian Witte
Publication Date: 2025-11-21
Creators: Maximilian Witte

This dataset consists of two complementary components capturing both the official positions of the major political parties in the 2021 German general election and the public perception of these positions.

The first component contains pre-processed short versions of the election programs from the six major parties that competed at the federal level in 2021: CDU/CSU, SPD, Bündnis 90/Die Grünen, FDP, Die Linke, and AfD. All statements are processed for text classification and were sourced from party publications released for the 2021 campaign.

The second component includes 7,500 individual statements from consumers describing what they believe these political parties stand for and do not stand for. Participants were asked to freely express their perceptions without constraints on length or structure. Each statement is linked to the referenced party where applicable and contains metadata on anonymized participant ID, timestamp, and language. Statements include both supportive and critical assessments and therefore represent a wide range of public interpretations of party identity and political priorities.

Together, the two components enable the study of the relationship between the official communication of political parties and the way citizens mentally represent party beliefs. The dataset can be used for research in political communication, perception gaps, misinformation, narrative framing, and natural language processing applications such as stance detection and text similarity.

Car Design Ratings for Text Analysis

Car Design Ratings for Text Analysis

Maximilian
Open Dataset
Creators: Maximilian Witte
Publication Date: 2025-11-21
Creators: Maximilian Witte

This dataset captures consumer evaluations of car design wireframes along three perceptual dimensions relevant to automotive styling research: aggressiveness, complexity, and typicality. It contains no visual material and is therefore designed exclusively for text-based analysis.

The dataset comprises 232 distinct car design wireframes, each represented through text descriptions detailing the visual form of the design. For every wireframe, consumers rated perceived aggressiveness, perceived complexity, and perceived typicality. Each dimension includes 20 independent ratings per wireframe, resulting in more than 13,000 numeric evaluations. The ratings are stored in separate CSV files, one for each dimension, and include the wireframe ID, anonymized rater ID, and the numeric score.

In addition to the numeric evaluations, a free-text description provided by participants accompanies every wireframe. These statements capture how consumers interpreted individual design elements and why they formed their perceptions. Together, the rating data and participant statements enable quantitative and qualitative analyses of design perception.

The dataset supports a wide range of applications including text-based modeling of aesthetic impressions, computational analysis of design language, semantic feature extraction, prediction of numeric perception ratings from text, and research on variability in consumer interpretations of automotive forms.

Datasets for Image Classification in Marketing

Datasets for Image Classification in Marketing

Maximilian
Open Dataset
Creators: Maximilian Witte
Publication Date: 2025-11-21
Creators: Maximilian Witte

This dataset provides a curated list of 18 publicly available image classification datasets selected for their relevance to marketing and consumer insight research. Each dataset contains labeled visual material that reflects consumer-facing content, product perception, or brand-related context. The collection covers a broad range of marketing-relevant domains including product packaging, retail environments, shelf placement, food and beverages, fashion, automotive, advertisements, and social media imagery.

Sign In

New to BERD@NFDI? Register

Forgot password?

Register

Reset Password

Please enter your username or email address, you will receive a link to create a new password via email.