Here is a summary of projects i have done for previous jobs, grouped by subject
Natural Language Processing
-
Classification model of digital news for national newspaper company, through feature engineering evaluating grammar rules and styles of writing using part of speech tagging and regular expressions.
-
Creation of use cases tags applying regular expressions and context knowledge for digital whiteboard company, taking as input titles of canvas files.
-
Topic modeling, word clouds, lemmatization for several exploratory data analysis and visualization deliverables.
Predictive Modeling
-
Model to assign probability of customer churn according to product usage.
-
Natural Language Processing and transfer learning to create a book genre classification system based on their titles and content description (written in spanish), for a company of book delivery.
-
User segmentation with behavioral variables and usage data extracted from open databases, experiments, observations and surveys to provide strategies to public transportation system company.
Ad hoc Analysis
-
Creation of score to classify canvas usage according to product behavior into synchronous and asynchronous, for digital whiteboard company.
-
Analysis to find frontend and backend events related with low Frames Per Second (FPS) performance in product.
-
Analysis of Net Promoter Score (NPS) to find variables related with changes and to track feedback of users.
-
Analysis of features in beta and post-launch stages to provide insights related with user adoption.
Visualization
-
Graphs for data exploration and visualization of results through Python packages like pyplot, seaborn and matplotlib.
-
Dashboards for stakeholders in Periscope Sisense and Tableau.
Automation/ETL
-
Data availability processes through Airflow (DAGs) and AWS (S3 for parquet storage, Glue crawlers).
-
Creation of automated reports in google sheets (BigQuery integration) and Microsoft Excel (macros in VBA).