LinkedIn Scraper + ESG NLP Scoring, 1.4M Posts (Client)
Tran Tien Van
Client project for a quantitative research team (client unnamed) that needed Stata-ready panel data on more than 300 companies' LinkedIn activity, with employee posts tagged by ESG content.
What we built
A five-phase scraper for company profiles, posts, engagements and the profiles of the people engaging
An ESG scoring pass with E, S and G dictionaries and N-gram matching
Delivery as Stata .dta panel files
Results
300+ companies tracked across multi-month windows
1,395,138 employee posts scored for E, S and G; 108,376 carry a score
Panel dataset of 146 MB+ delivered as Stata .dta files
What the data says
The scorer is deliberately conservative: 7.8% of posts get a non-zero score
Social dominates; environmental is the rarest signal
Post volume peaked in 2021 and has since turned down
The dashboard was built by Van Data Team for its portfolio from the delivered data; the client engagement delivered the data.
Stack: Python, web scraping, NLP (dictionaries + N-grams), Stata.
Full case study: vandatateam.com/portfolio
Like this project
Posted Nov 5, 2026
Client project: five-phase LinkedIn scraper for 300+ companies, 1.4M employee posts scored for E, S and G with NLP, delivered as Stata panel data.
Likes
0
Views
0
