Big Data Days 2019

 October 8-10   Moscow

Valentina Djordjevic

Things Solver, Serbia


Valentina Djordjević is a passionate Data Science enthusiast. She works as a Data Scientist at Things Solver, since October 2016. She has a Bachelor’s degree in Information Systems and Technologies and a Master’s degree in Business intelligence, at University of Belgrade, Faculty of Organizational Sciences. The main fields of studies she focuses on the most are time series analysis and anomaly detection techniques. She has a strong technical knowledge in the field of Data Science, including programming (Java, Python, SQL, ETL), statistics (descriptive statistics, hypothesis testing, probability theory,…), modeling (machine learning algorithms – ensemble-based, neural networks, similarity-based, …) and visualization (Matplotlib, Seaborn, Plotly, Tableau,…). She’s working on various Data Science problems coming from different business domains, from telecommunications to retail, finance and banking, where she’s dealing with forecasting, predictive maintenance, anomaly detection, segmentation, churn prediction, lead generation and scoring, etc.


Breaking Out the Lead Scoring Algorithm

The Lead is considered to be any individual who may become a potential client as it has shown an interest in the product or service a company offers.
A Lead Generation refers to the method of collecting leads in order to manage sales channels more efficiently, raise brand awareness and contribute to rising profits. Lead Scoring includes assigning certain weights to each potential lead activity, thereby ranking their behavior according to the likelihood that they may be converted into customers.
This topic will show how Data Science can be used in order to optimize the process of lead analysis. The presentation will cover a real-world example, where advanced analytics showed that even having a perfect lead scoring algorithm is not enough to increase the number of converted leads. The main idea is to present the benefits of using Data Science to extract insights, detect patterns of behavior and identify bottlenecks in the process of lead generation, in order to optimize it and increase conversion rates.

Session Keywords

Lead Scoring
Data Science


Machine Learning Techniques for Anomaly Detection

This workshop is dedicated to the machine learning techniques that can be used for anomaly detection. The session is organized in three phases, where each phase is more advanced and demanding than the previous one.

Phase 1 includes theoretical anomaly detection introduction and covering basic anomaly detection techniques like z-score and smoothed z-score for anomaly detection.

Phase 2 includes more advanced machine learning algorithms, able to work with multivariate datasets, like Isolation Forest and Elliptic Envelope.

Phase 3 includes using the Autoencoder neural network to detect anomalies in huge multivariate datasets.