Global asset manager (AuM: USD700bn+): Natural language processing for ESG investing

  •  30

    data fields analysed

  • 10

    alternative datasets sourced

  • 4

    NLP engines built


CLIENT PROFILE AND CHALLENGES
  • Director of ESG Integration and Research of a global institutional asset management firm
  • The client wanted an in-house ESG scoring framework that was fully transparent, able to provide real-time updates and customisable to the current portfolio
  • Scope included analysing textual data sourced from news/company publications to create sentiment scores based on UN SDGs
OUR APPROACH
  • Deployed a three-member team – 1 lead data scientist, 1 NLP specialist and 1 data engineer – to build, test, deploy and fine-tune the NLP engines
  • Built a classification model to identify SDG statements within company reports
  • Build a bespoke sentiment analysis model for news data
  • Created a bespoke UI for portfolio managers

IMPACT DELIVERED

  • A proprietary in-house ESG scoring framework based on NLP engines – transparent, able to provide real-time updates and scalable
  • Algorithms to power NLP engines with overall accuracy of 85% and full ownership of intellectual property
  • 70%+ operational cost savings by leveraging a skilled data science team working from an offshore delivery centre
Thank you for sharing your details

Your file will start downloading automatically

If it does not download within 1 minute,

Share this on

What we have done

Data tagging and analysis focused on fundamental sub-sector equity strategy
What we are proud of

800+ merchant names tagged

Optimised and scalable solution across datasets and sectors

Enterprise-level data management powered by analytics
What we are proud of

Expertise in market data

Handled end-to-end data-pipeline transformation for more than 1,000 datasets

US asset manager: Non-linear factor modelling for systematic investment strategies
What we are proud of

10,000+

assets analysed

5+

external data pipelines built

US private credit fund: Automated and self-service analytics for enhanced portfolio insights
What we are proud of

100+

funds

6

strategies

Leading South African bank: Alternative data analysis to differentiate equity research
What we are proud of

500m+

real-estate property data points

4

websites

API-based access to analyst credit research model data for advanced analytics and visualisation for top-tier European asset manager
What we are proud of

100%

Automation in the publishing of model data

Improved data quality

Risk data, application management and reporting for a consortium of 20 development banks
What we are proud of

100%

automated report generation process

40%

Reduced cost of ca.

CRO risk data, CECL, analytics and reporting for a US-based bank
What we are proud of

Offered scalable automated solutions

Centralised one data warehouse