Large-scale Processing of Social Media Data

In this talk, we will summarize our experience with processing social media data on a large scale. We will give an overview of most of the tasks related to mining insights from social media data. In Socialbakers (now Emplifi), we have been processing terabytes of social media data every week over the last 10 years. We want to share our experience from building various ML models, from NLP tasks (such as sentiment analysis, NER, or NLG systems) to computer vision applications (such as object detections) to time series analysis for crisis management.

We will introduce our AI-based data enrichments. These are various NLP and CV methods we use to transform and structure unstructured data from social networks. We use these enrichments as building blocks and combine them to get more relevant insights for our clients. On a higher level, we will discuss tools and features for social media marketers, such as influencers management, social media listening, and AI-based features that help marketers manage interactions with their clients.

The talk is aimed at data scientists and data analysts, but in general at anyone interested in learning about social media data, obtaining valuable information, and building on top of it.

About the speaker

Peter Krejzl

Director of Research at

Peter Krejzl is a Director of Research at, a unified CX platform for enterprises. Peter manages the international research team working on a wide variety of machine learning solutions, from natural language processing to image analysis, anomaly detection, or conversational AI.

The team helps apply machine learning to solve tasks across the whole unified CX platform including marketing, care, and commerce solutions. Peter’s main point of interest is natural language processing and applications of AI in various marketing and care tasks.



Sessions: October 5 – 7
Trainings: October 4, 12 – 15


Presented by