Portfolio
John Bianchi

As part of the FAO-ASFA project, ISTI-CNR is engaged in the implementation of a text mining service for documents on fisheries, aquaculture and marine biodiversity. During the proposed internship, it was required to implement an algorithm for extracting categorized information from text, exploiting the Web service named NLPHub of ISTI-CNR. The algorithm was finally published in as-a-service mode on the CNR cloud computing platform named DataMiner, using the software import tools offered by the platform itself.

The work was carried out in collaboration with Dr. Gianpaolo Coro who led the project.


DataMiner service webpage

GitHub repository

Documentation (available only in Italian)

Full-Stack e-commerce website development using Salesforce commerce cloud

E-commerce link

Report realized for the course of Data Mining: Fundamentals based on the dataset containing vehicle data belonging to Carvana Co.

Documentation

Final essay developed for the course "Artificial Intelligence & Society" at King's College London.

"This paper tries to put itself as a point of reflection on what is the discipline of machine learning extending the argumentation also towards the main companies that use this artificial intelligence technology. The lack of ethical attention regarding the content of training datasets is an intellectual weakness that should not be underestimated in order not only to safeguard the environment and prevent forms of discrimination but also to establish a conscious and genuine relationship with this important form of technology."

The price of automatic communication

Website developed in collaboration with Fabio Camillo on request of Wally Communication

Bagno Imperia website

Website created on request of Wally Communication

El Niño website

Web scraping of Tripadvisor using a Raspberry Pi 3 Model B+ and development of a dedicated website for the visualization of the extracted data.
This project constitutes the thesis of my bachelor's degree in digital humanities.


Documentation (available only in Italian)

Development of a machine learning model that predicts which Tweets are about real disasters and which are not.

Created for the "Text Analytics" exam, this project illustrates the realization of two different models both able to operate this classification task: the first one, based on a traditional approach, uses the technique of feature engineering while the second one uses the Bert language model through Simple Transformers library.

The work was carried out in collaboration with Dr. Andrea Esuli who has oversaw the project.


Overview of project

Project Repository

This archive lets you explore the music albums through the colour of their covers. You can enjoy it by randomly exploring the visualizations or by looking at the specific categories. All the music data, including images, are provided by Spotify. The authors' contribution is relative to the chromatic analysis of the data and to their visual and logical organisation.

Website developed in collaboration with Marta Fioravanti

Link to website