We provide a 360° assessment of corporate ESG performance, by combining internal data (traditional, self-disclosed data) with external ‘alternative’ data (stakeholder-generated data) in order to measure the gap between what the companies communicate and what is stakeholder perception related to corporate sustainability commitments.
Dedicated to ESG Management, Investor Relations, Asset Management, Credit & Risk Management
ESG performance and value creation.
A recent study* revealed that companies with a high ESG rating performed better in terms of budgetary outcomes and stock returns. As a consequence, there is a need for new models of ESG performance analysis, which could help companies to manage risks and opportunities and to increase profit and value on the long term.
*Banor SIM in partnership with the Politecnico di Milano; data refer to 882 stocks comprised in the Stoxx Europe 600 Index in the period 2010-2017.
External Alternative Data as measuring factor.
Measuring ESG performance only through traditional in-bound information (data published by the company itself) could lead to an incomplete assessment, as data is partial and the nature of such disclosure is strictly connected to the particular reporting framework chosen by the company.
Therefore integrating such assessment with the analysis of great amounts of external Alternative Data – by adopting an AI-based approach – generated by company stakeholders provides a more complete view of corporate ESG performance in terms of reputation.
Tracking corporate contribution to UN SDGs
An ESG solution to support:
- 400+ SDG-related indicators
- 100,000+ data sources
- AI-driven data collection
- Weekly updates
- Sentiment-analysis integration
- Corporate reputation evaluation
- Green/social washing detecting tools
- Controversial activities detection
- Benchmarking tools
FinScience daily retrieves news content from the web. About 1.5 million web pages are visited every day on about 35.000 different domains. The content of these pages is extracted, interpreted and analysed to identify valuable information and sources. The FinScience’s news data pipeline can be divided into 3 steps, depicted below:
This phase involves the collection of data from different web sources: websites, social network pages, news or blogs. The latter are identified mainly by following two different criteria that are not necessarily connected to each other: (i) the level of sharing and visibility of a content; (ii) the identification of specific sources linked to a particular topic of interest.
Then, the contents are extracted from the web pages and are pre-processed via a first level of data cleaning for the elimination of noise and the extraction of the main body: this is the input to the next phase of data processing.
At this stage, Natural Language Processing (NLP) methods are carried out. The contents collected in the data gathering phase are subjected to an NLP analysis that allows to determine the objects (companies and SDG topics…) disseminated and discussed on the web.
Data analysis and enrichment
Once the topics covered have been identified, the data are analysed, normalised and enriched to obtain further metrics such as:
Digital popularity Value (DPV): a measure of the diffusion of a digital signal on the web. It is obtained by aggregating the diffusion metrics of the news mentioning the signal at hand and can take only positive values.
Sentiment: it measures how users feel about a specific company or information and can take values in the interval (-1, 1). This indicator seeks to quantify how current beliefs and positions affect future behaviors. Our sentiment analysis algorithm is based on a semi-supervised method.
The sentiment value – weighted by DPV – is finally assigned for the company in relation to each SDG to reflect the public perception of corporate sustainability efforts/impact.
The Scoring process
Data are automatically collected from the different sources described below with different frequency (daily, weekly or annually).
Conversion into SDG-related indicators
Textual content is first analyzed via Natural Language Processing tools such as text classification, entity extraction and sentiment analysis.
SDG Meta-Score calculation
The data sources are classified into internal or external, according to whether or not it is voluntarily disclosed by the company. Internal and external scores for each SDG are computed.
Internal/External Score calculation
External and internal scores are obtained by aggregating the SDG scores, also taking into account the industry in which the company operates by means of materiality matrix.
Final FinScience ESG Score
The final FinScience ESG score is obtained by averaging internal and external scores, applying a penalty coefficient for those companies where internal and external disclosure results as conflicting.
Benefits from our integrated and modular approach:
To learn more about the use of Alternative Data for companies’ ESG assessment, become a free member of this LinkedIn group.
The added-value of Alternative Data for ESG assessment
Find out more about FinScience’s services and solutions most suited to your business.