ADM_HW4

Repository for the 4th homework of the course ADM @ Sapienza University of Rome from group group #23 composed by Francisca Alliende, Giuseppe Calabrese and Francesco Russo

Incoming, a summary of the files of this repository. To access to a document just press the link in the name of the corresponding file.

Homework_4

Jupiter Notebook, with the code and coments of the entire homework

First Part: Does basic house information reflect house's description?

Modules:

• scraper.py: functions related to the scraping process.
• preprocessing.py: functions related to the preprocesing process.
• matrixbuilder.py: functions that build the information and the description matrices.
• clustering.py: functions of k-means++, Elbow Method and Jaccard Similarity.
• wordcloudgenerator.py: wordcloud generator function.
• mykmeans.py: homemade k-means algorithm.

Databases:

• datasetindex.csv: database with all the announcements after the scrapping process.
• datasetIndex_preprocessed.csv: database that contains the data from "datasetindex.csv", prepocessed.
• datastIndex_infmatrix.csv: database with the informatrion matrix. Input for clusterization.
• datastIndex_tfidf.csv: database, with the description matrix. Input for clusterization. Unfortunately not available due to its weight.

Second Part: Find the duplicates!

All the code and comments of this parts, they are contained in the main file Homework_4