Demian Pablo

Working on Data Science projects for GSK (Glaxo Smith Kline) account for US Market.
Using AWS instances for connecting to a set of different data sources (Experian, Comscore, IRI, Epsilon, Google DV360), mostly by using Pyspark to perform data manipulation, feature engineering and several types of modeling approaches (such as Logistic Regression, Random Forests, GBT, MLP, Naive Bayes, Linear Support Vector Machines, etc).

Projects include EDA, audience profiling, time lag analysis, look alike modeling, among others.
Working in collaboration with Strategy and Media team, providing insights for increasing the effectiveness of content marketing and activation.
Development of a deep understanding of target audiences through exploratory analysis to generate hypotheses and identify growth opportunities.
Building scalable code for working with very large data sets; and making it flexible, in order to be leveraged by other members across the team.

Leading Data Science practice in Argentina.
Working on audience data collected via Amazon Redshift from a set of DMPs, and also from clickstream data providers (such as Lotame and Jumpshot). Tableau dashboards for data visualization.
Audience profile setting and consumer journey taxonomy definition by working along with the Strategy team.

Web scraping from blogs and Youtube comments (by using Python’s Beautiful Soup and Selenium packages)
Natural Language Processing of text data for spanish language; which consist in tokenization, use of regular expressions, stopwords, and tailor made functions in Python.
Creation of custom sentiment classifier from scratch on Jupyter notebooks by using bag of words and Hamming distance in order to get classes (positive, negative and neutral) for each unlabeled comment scraped, and then refined by the implementation of n-grams and k-nearest neighbors.

Responsible for delivering marketing insights across media touchpoints and audiences to drive the strategic planning process, evolve measurement and optimize the media mix to improve return on investment.
Local project manager for data integration from a variety of digital sources (such as Facebook and Google), which consists in extract, transform and load that data in order to automatize performance reports and provide meaningful insights across all teams and clients.
Deployment of agency optimization tools designed to improve marketing insights and operations (mostly for predictive analytics, campaign performance and KPIs visualization).
Collaboration across teams for producing strong insights to reporting and client deliveries.

Business strategies definition.
Set different scenarios. Revenue, costs and profit forecast for diverse markets.
KPIs analysis for each business unit (Branding, Apps, VAS). Trends determination.
Integration of data from diverse platforms (Google, Facebook, Appnexus, Double Click, Tube Mogul, etc) for reporting.
Working with IT in order to implement/improve processes.
Klipfolio’s Admin. In charge of building dashboards for data visualization.

System users administrator of CRM software. Accounts creation (digital advertising agencies), commercial opportunities/proposals flow management. In charge of user´s training.
Revenue and strategic reports (accounts categorization, sales activity measurement, coverage territory, targets by agencies and sales reps, win rates by product, sales achievement).
Benchmarking (Competitive market analysis).
Database manager for Marketing campaigns.

In charge of post-selling service. Management of goods supply, constant contact with Leasing´s customers and suppliers.
Generation of Leasing contracts by using the Oracle system.
Control of financial contributions, rates and levies for leasing operations.

End to end data pipelines.
NLP and Sentiment Analysis.
Automation of tasks working with different APIs, Xlwings & GCP Cloud Functions.
Statistical modelling and machine learning
Segment audiences using algorithms such as K-means or PCA.
Development of performant, scalable and reusable production code, working in collaboration with other team members through GitHub.

ETL pipelines.
Feature Engineering.
Unit Testing & code reviews.
Code refactoring & debugging.
Building of R packages.
Shiny dashboards.
Time Series Analysis & Gating Checks.
Stack: R, R Shiny, Scala, GIT, Bitbucket, Azure, Docker.
Client: Bridgewater Associates (Finance/Trading).