Login
- Dokumenttyp:
- Masterarbeit
- Autor(en):
- Geske, Flora (FIM)
- Titel:
- Social-Media based NLP-powered ESG Scoring Methodology
- Abstract:
- With the fast-growing attention society pays to the ESG-impact of companies, and with investors being more diligent with respect to sustainable investing, it is crucial to provide objective and accurate ESG information about companies. The aws traditional ESG scores su_er from, ranging from methodological weaknesses to systematic skews, are causing a mis-reection of ESG issues in the market and do not guide companies to improve their actions. While the industry has taken initial steps towards innovating ESG ratings, leveraging technology and alternative data sources, the research domain is still in its infancy to include such approaches. This thesis follows a two-fold objective: Addressing the aws of traditional ESG scoring as well as establishing a framework for a technology-powered, stakeholder data-based ESG score methodology in the research domain. Our framework forms the basis for a dynamically updating, objective, independent, and e_cient evaluation of _rms with respect to ESG criteria. We build an ESG scoring mechanism using social media and Natural Language Processing (NLP) technology to provide proof of concept. We evaluate our approach with three case studies and a sample of 175 _rms from the S&P 500 index, showing that we can measure materially relevant ESG information voiced by stakeholders and successfully address the aws of traditional ESG scoring.
- Aufgabensteller:
- Prof. Zagst, Prof. Luis Seco, University of Toronto
- Betreuer:
- Jonathan Mostovoy
- Jahr:
- 2020
- Hinweise:
- Masterarbeit für FIM
- Hochschule / Universität:
- Technische Universität München
- TUM Einrichtung:
- Lehrstuhl für Finanzmathematik
- Bearbeitungsbeginn:
- 01.12.2020
- Bearbeitungsende:
- 31.05.2021
BibTeX