Paper 1: Real Estate Media Sentiment Through Textual Analysis
This is the first paper to capture media sentiment from news that is relevant to U.S. securitized real estate markets. For this purpose, 125,000 newspaper article headlines from four different sources, namely Bloomberg, The Financial times, Forbes and The Wall Street Journal, were collected. Subsequently applying the dictionary-based approach and aggregating the results, led to three different sentiment measures on a monthly level. In a vector autoregressive framework, the created sentiment measures yielded a significant relationship with future REIT market movements.

Paper 2: News-Based Sentiment Analysis in Real Estate: A Machine-Learning Approach Via Support Vector Networks
This study applies a machine-learning approach, namely support vector networks, to capture sentiment from professional news headlines published by S&P Global Market Intelligence database (SNL). For the first time, sentiment measures based on a support vector machine as a classification algorithm, are investigated and compared regarding to their relationship with the U.S. securitized and direct commercial real estate market.

Paper 3: Predicting Real Estate Market Movements: the First Textual Analysis-Based-Sentiment Application in Germany
The major aim of this research paper is to lay the foundation for and test the applicability of text-based sentiment analysis in German real estate markets. First, the groundwork was accomplished by developing the first German Real Estate Sentiment Dictionary. The next steps yielded robust evidence of a significant relationship between the extracted negative sentiment measure and the residential real estate market in Germany.

Die Dissertation untersucht den Einfluss von Markt-Stimmung, extrahiert aus immobilienbezogenen Zeitungsartikeln, auf die amerikanischen und deutschen Immobilienmärkte. Dabei werden innovative Ansätz der Sentiment Analyse verfolgt, die es ermöglichen Stimmung zu quantifizieren. Durch die sogenannte Textanalyse werden Immobilien-Nachrichten aus verschiedenen Quellen analysiert und als ein monatliches Stimmungs-Maß aggregiert. Die Ergebnisse weisen auf signifikante Zusammenhänge zwischen den kreierten Stimmungs-Indikatoren und den zukünftigen Entwicklungen auf den Immobilienmärkten hin.