Im a MSc student in Data Science at the University of Pisa. I am a curious, inclusive and determined person with a great desire to learn more and more and achieve my personal goals. Address: Italy
Project given by the "Text Analytics" course of the University of Pisa. The project was conducted... more Project given by the "Text Analytics" course of the University of Pisa. The project was conducted in a team of 3 people, we deep dive into it with different algorithms/experiments to find the optimal solution to the problem along with the comparison of multiple models.
The dataset used for this project is obtained from the Kaggle competition “Quora Question Pairs” and it is openly available on the website Kaggle.com. In our approach we implemented many classical machine learning models like logistic regressions, random forest and so on. Subsequently we configured BERT transformers with different model type.
To run the code which we enclose in the submission, we used the Kaggle platform in order to get the most out of GPU Accelerator.
ETL Project given by the University of Pisa conducted in a team of 3 people. We worked hard on th... more ETL Project given by the University of Pisa conducted in a team of 3 people. We worked hard on this laboratory project to develop a Data Warehouse and explore it using SQL. Therefore we deployed a Cube and explored it with MDX before creating dashboards through Microsoft Power BI
The tools used for this project are: Microsoft Visual Studio, Microsoft SQL Server and Power BI.
In particular I worked on building and populating the Data Warehouse, manipulating data in formatc csv using Python.
Caritas Lodigiana "Da sempre vicino agli ultimi", 2021
Project organized in collaboration between the Master's course "Technologies for web marketing" a... more Project organized in collaboration between the Master's course "Technologies for web marketing" at the University of Pisa and Google Marketing Challenge. This project consisted of a 1 month Google Campaign for a non-profit association, in our case the Caritas Lodigiana Association. It allowed us to develop and put into practice the different Google Tools, involving: - Google Ads - Google Analytics - Google Search Console - Google Tag Manager - Google Optimize
Data Science project given by the University of Pisa conducted in a team of 4 people.
We try an a... more Data Science project given by the University of Pisa conducted in a team of 4 people. We try an analysis of the Carvana dataset with Data Mining tools. The goal is to predict whether a vehicle is a good buy or not. In the first section the data have been explored, in order to evaluate and improve the data quality (semantic errors, redundancies, inconsisten-cies and outliers) and make some statistical considerations about the target problem(good buy or bad buy). In the second section, different clustering algorithms such as partition clustering with K-Means, density-based clustering with DBSCAN and hierarchical agglomerative clustering have beenused in order to explore the data and toassess the similarity between the records.In the third section, frequent patterns and association rules have been extracted with Apriori algorithm. In the last section, a model based on Decision Tree and Random Forest has been developed in order to predict the target variable.
Project given by the "Text Analytics" course of the University of Pisa. The project was conducted... more Project given by the "Text Analytics" course of the University of Pisa. The project was conducted in a team of 3 people, we deep dive into it with different algorithms/experiments to find the optimal solution to the problem along with the comparison of multiple models.
The dataset used for this project is obtained from the Kaggle competition “Quora Question Pairs” and it is openly available on the website Kaggle.com. In our approach we implemented many classical machine learning models like logistic regressions, random forest and so on. Subsequently we configured BERT transformers with different model type.
To run the code which we enclose in the submission, we used the Kaggle platform in order to get the most out of GPU Accelerator.
ETL Project given by the University of Pisa conducted in a team of 3 people. We worked hard on th... more ETL Project given by the University of Pisa conducted in a team of 3 people. We worked hard on this laboratory project to develop a Data Warehouse and explore it using SQL. Therefore we deployed a Cube and explored it with MDX before creating dashboards through Microsoft Power BI
The tools used for this project are: Microsoft Visual Studio, Microsoft SQL Server and Power BI.
In particular I worked on building and populating the Data Warehouse, manipulating data in formatc csv using Python.
Caritas Lodigiana "Da sempre vicino agli ultimi", 2021
Project organized in collaboration between the Master's course "Technologies for web marketing" a... more Project organized in collaboration between the Master's course "Technologies for web marketing" at the University of Pisa and Google Marketing Challenge. This project consisted of a 1 month Google Campaign for a non-profit association, in our case the Caritas Lodigiana Association. It allowed us to develop and put into practice the different Google Tools, involving: - Google Ads - Google Analytics - Google Search Console - Google Tag Manager - Google Optimize
Data Science project given by the University of Pisa conducted in a team of 4 people.
We try an a... more Data Science project given by the University of Pisa conducted in a team of 4 people. We try an analysis of the Carvana dataset with Data Mining tools. The goal is to predict whether a vehicle is a good buy or not. In the first section the data have been explored, in order to evaluate and improve the data quality (semantic errors, redundancies, inconsisten-cies and outliers) and make some statistical considerations about the target problem(good buy or bad buy). In the second section, different clustering algorithms such as partition clustering with K-Means, density-based clustering with DBSCAN and hierarchical agglomerative clustering have beenused in order to explore the data and toassess the similarity between the records.In the third section, frequent patterns and association rules have been extracted with Apriori algorithm. In the last section, a model based on Decision Tree and Random Forest has been developed in order to predict the target variable.
Uploads
The dataset used for this project is obtained from the Kaggle competition “Quora Question Pairs” and it is openly available on the website Kaggle.com. In our approach we implemented many classical machine learning models like logistic regressions, random forest and so on. Subsequently we configured BERT transformers with different model type.
To run the code which we enclose in the submission, we used the Kaggle platform in order to get the most out of GPU Accelerator.
The tools used for this project are: Microsoft Visual Studio, Microsoft SQL Server and Power BI.
In particular I worked on building and populating the Data Warehouse, manipulating data in formatc csv using Python.
- Google Ads
- Google Analytics
- Google Search Console
- Google Tag Manager
- Google Optimize
We try an analysis of the Carvana dataset with Data Mining tools. The goal is to predict whether a vehicle is a good buy or not. In the first section the data have been explored, in order to evaluate and improve the data quality (semantic errors, redundancies, inconsisten-cies and outliers) and make some statistical considerations about the target problem(good buy or bad buy). In the second section, different clustering algorithms such as partition clustering with K-Means, density-based clustering with DBSCAN and hierarchical agglomerative clustering have beenused in order to explore the data and toassess the similarity between the records.In the third section, frequent patterns and association rules have been extracted with Apriori algorithm. In the last section, a model based on Decision Tree and Random Forest has been developed in order to predict the target variable.
The dataset used for this project is obtained from the Kaggle competition “Quora Question Pairs” and it is openly available on the website Kaggle.com. In our approach we implemented many classical machine learning models like logistic regressions, random forest and so on. Subsequently we configured BERT transformers with different model type.
To run the code which we enclose in the submission, we used the Kaggle platform in order to get the most out of GPU Accelerator.
The tools used for this project are: Microsoft Visual Studio, Microsoft SQL Server and Power BI.
In particular I worked on building and populating the Data Warehouse, manipulating data in formatc csv using Python.
- Google Ads
- Google Analytics
- Google Search Console
- Google Tag Manager
- Google Optimize
We try an analysis of the Carvana dataset with Data Mining tools. The goal is to predict whether a vehicle is a good buy or not. In the first section the data have been explored, in order to evaluate and improve the data quality (semantic errors, redundancies, inconsisten-cies and outliers) and make some statistical considerations about the target problem(good buy or bad buy). In the second section, different clustering algorithms such as partition clustering with K-Means, density-based clustering with DBSCAN and hierarchical agglomerative clustering have beenused in order to explore the data and toassess the similarity between the records.In the third section, frequent patterns and association rules have been extracted with Apriori algorithm. In the last section, a model based on Decision Tree and Random Forest has been developed in order to predict the target variable.