Enrique J. D.

Enrique J. D.

Data Scientist

Barranquilla , Colombia

Experience: 15 Years

Enrique J.

Barranquilla , Colombia

Data Scientist

30000 USD / Year

  • Immediate: Available

15 Years

Now you can Instantly Chat with Enrique J.!

About Me

Im an Industrial Engineer turned Data Scientist with more than fifteen years of experience in the industry and academic sector. Im currently a lecturer of undergraduate and graduate courses in Data mining, Big Data, and Machine Learning. Im fluent in...

Show More

Portfolio Projects

Description

To foster the user experience in the virtual drugstore "tudrogueriavirtual.com", we developed a recommendation algorithm considering the legal restrictions regarding the sale of prescription drugs.

- Using Content-Based Filtering, User-BasedFiltering, and association rules, I led the creation of four market segments, representing customer-driven categoriesused for recommendations.

- Build a predictive model using Random Forest for estimating customer loyalty scores (in construction)

- The model was put into production in AWS, using Docker and Sagemaker.

Show More Show Less

Description

- In 2020 Movistar decided that all mobile phone plans would become unlimited.

- To assess the uncertainty of the decision, I built a Time Series model, bringing confidence intervals from 0 - 360 years in advance.

- The link provided is just a short example of the model. For legal reasons, I can't provide to you with the whole development.

Show More Show Less

Description

- Sagicc is an omnichannel platform that helps companies interact with their clients efficiently.

- I built a recommender system to determine the best channel to set communications with clients and maximizecustomer satisfaction.

Show More Show Less

Description

-Power theft is a serious problem in Colombia. Thus, in association with the Electricaribe company, a pilot model was designed to predict energy theft using a database of 78,712 users in the city of Barranquilla.

- Three feature selection methods were used to reduce the number of final predictors. The models were evaluated through the accuracy and the F1 score using a 10-fold cross-validation algorithm. Results showed that the final subsets provided enough overall performance. However, the best subset correspond to the Tree-based subset. A gradient boosting machine was the model outperformed the rest, giving a mean accuracy of 74.3% and an F1 score of 83.1. These results represent great insights to local DSO and utilities to join artificial intelligence to their energy meters to reduce NTL significantly and therefore increase profit and reliability.

Show More Show Less

Description

-Target emotions were elicited in 37 volunteers using video clips while two biosignals were recorded: photoplethysmography, which provides information about heart rate, and galvanic skin response.

-Several feature selection techniques and classifiers were evaluated. The best model was obtained with random forest recursive feature elimination, for feature selection, and a support vector machine for classification.

- This project was publicated in theBiomedical Signal Processing and Control journal.

Show More Show Less

Description

-The disappearance of people is a multidimensional phenomenon, in which several aspects must be considered. It affects people’s security perception and consumes police resources in its treatment.

- In association with the General Prosecutor’s office of Cartagena - Colombia, we developed a model to predict the future state of a missing person.

-We built three supervised machine-learning algorithms, K-Nearest Neighbours, Decision Trees, and Random Forest.

Show More Show Less

Description

- We built a robust model to analyze the urban mobility in Bogota City.

- We built aforecast model based on Random Forest to predictthe travel time of a citizen, considering the motive and time of the travel, destination, origin and other related variables.

Show More Show Less

Description

- To combat the no availability of parking positions, we developed a simulation model in R. Considering the volume production (500.000 cars per year, meaning 1 car per minute), we set two objectives, reduce the distance to parking and the number of movements.
- As a result, two new parking rules were created, reducing in 13% the travelled distance overall, and passing from 3.1to 2.8 movements per vehicle.

Show More Show Less

Description

Activity: Experimental design for validating the technological solution.

Show More Show Less

Description

A/B Experiment to validate the technologica solution.

Show More Show Less

Description

Activity: Design data colletion process and machine learning modeling at the Cerrejón mine in irrigation tanker trucks.

Show More Show Less

Description

Show More Show Less

Description

Design a recommender system based on Machine learning algorithms

Show More Show Less

Description

Experimental design for validating the technological solution.

Show More Show Less

Description

Data Analysis and machine learnig modeling.

Show More Show Less

Description

Design data colletion process and machine learning modeling at the Cerrejón mine in irrigation tanker trucks.

Show More Show Less

Description

Projection of the aggregate demand after the widespread implementation of unlimited data plans in 2020.

Show More Show Less

Description

Project director.

Show More Show Less