Resources by Stefan

goodbooks-10k

goodbooks-10k

Stefan
Open Dataset
Creators: Zając, Zygmunt
Publication Date: 2017
Creators: Zając, Zygmunt

The dataset contains six million ratings for ten thousand most popular books (with most ratings). It offers a rich resource for analyzing reading habits, book popularity, and user engagement within the literary community. There are also books marked to read by the users, book metadata (author, year, etc.) and tags/shelves/genres.

ratings contains ratings sorted by time. Ratings go from one to five. Both book IDs and user IDs are contiguous. For books, they are 1-10000, for users, 1-53424.

to_read  provides IDs of the books marked “to read” by each user, as user_id,book_id pairs, sorted by time. There are close to a million pairs.

books has metadata for each book (goodreads IDs, authors, title, average rating, etc.). The metadata have been extracted from goodreads XML files.

book_tags contains tags/shelves/genres assigned by users to books. Tags in this file are represented by their IDs. They are sorted by goodreads_book_id  ascending and count descending.

The date set is 68.8 MB large.

European Funds dataset from Morningstar

European Funds dataset from Morningstar

Stefan
Open Dataset
Creators: (Leone, Stefano)
Publication Date: 2019
Creators: (Leone, Stefano)

The file contains 57,603 Mutual Funds and 9,495 ETFs with general aspects (as Total Net Assets, management company and size), portfolio indicators (as cash, stocks, bonds, and sectors), returns (as yeartodate, 2020-11) and financial ratios (as price/earning, Treynor and Sharpe ratios, alpha, and beta).
Additional data in terms of sustainability is also available. A key feature of this dataset is the inclusion of detailed Morningstar ratings, which are widely used in the financial industry to assess fund quality based on past performance, risk-adjusted returns, and analyst evaluations. Additionally, it offers categorization of funds, allowing for segmentation by investment type, sector, region, and fund style (e.g., growth vs. value investing). The dataset has a total size of approximately 103.88 MB.

Overall, the dataset is structured into the following variables:​

  • ticker: Fund ticker code.​
  • isin: Fund ISIN code.​
  • fund_name: Extended name of the fund.​
  • inception_date: Date of the fund’s inception.​
  • category: Fund category.​
  • rating: Morningstar rating.​
  • analyst_rating: Morningstar analyst rating.​
  • risk_rating: Morningstar risk rating.​
  • performance_rating: Morningstar performance rating.​
TripAdvisor European restaurants

TripAdvisor European restaurants

Stefan
Open Dataset
Creators: (Leone, Stefano)
Publication Date: 2021
Creators: (Leone, Stefano)

TripAdvisor is the most popular travel website and it stores data for almost all restaurants, showing locations (even latitude and longitude coordinates), restaurant descriptions, user ratings and reviews, and many more aspects. The dataset is 0.68 GB large.

The TripAdvisor dataset includes 1,083,397 restaurants with attributes such as location data, average rating, number of reviews, open hours, cuisine types, awards, etc.

The dataset combines the restaurants from the main European countries, the data has been scraped in early May 2021.

The dataset is structured with various variables for each restaurant, such as:​

  • restaurant_link: Unique TripAdvisor restaurant link.​
  • restaurant_name: Name of the restaurant on TripAdvisor.​
  • original_location: Original location displayed on TripAdvisor.​
  • country: Country name retrieved from original_location.​
  • region: Region name retrieved from original_location.​
  • province: Province name retrieved from original_location.​
  • city: City name retrieved from original_location.​
  • address: Address displayed on TripAdvisor.​
  • latitude: Latitude coordinate.​
  • longitude: Longitude coordinate.​
  • claimed: Indicates if the restaurant business is claimed on TripAdvisor.​
  • awards: Award names.​
  • popularity_detailed: Detailed popularity ranking.​
  • popularity_generic: Generic popularity ranking (among all places to eat in the area).
Credit Card Fraud Detection

Credit Card Fraud Detection

Stefan
Open Dataset
Creators: Worldline and the Machine Learning Group of ULB ((Universite Libre de Bruxelles)
Publication Date: 2016
Creators: Worldline and the Machine Learning Group of ULB ((Universite Libre de Bruxelles)

The Credit Card Fraud Detection dataset is a rich collection of credit card transactions made by European cardholders in September 2013. It presents transactions that occurred in two days, where we have 492 frauds out of 284,807 transactions. The dataset is highly unbalanced, the positive class (frauds) account for 0.172% of all transactions. The dataset is 0.15 GB large.

The data has been collected and analysed during a research collaboration of Worldline and the Machine Learning Group (http://mlg.ulb.ac.be) of ULB (Université Libre de Bruxelles) on big data mining and fraud detection.

Each transaction record in the dataset includes several features:​

  • Time: The number of seconds elapsed between this transaction and the first transaction in the dataset.​

  • V1 to V28: These are the result of a Principal Component Analysis (PCA) transformation applied to the original features to protect sensitive information.​

  • Amount: The monetary value of the transaction.​

  • Class: A binary indicator where ‘1’ signifies a fraudulent transaction and ‘0’ denotes a legitimate one.​

Heart Disease Data Set

Heart Disease Data Set

Stefan
Open Dataset
Creators: Janosi, Andras; Steinbrunn, William; Pfisterer, Matthias; Detrano, Robert
Publication Date: 1988
Creators: Janosi, Andras; Steinbrunn, William; Pfisterer, Matthias; Detrano, Robert

The Heart Disease database is a well-regarded resource in the medical research community, particularly for studies related to cardiovascular conditions. It comprises data from four distinct databases: the Cleveland Clinic Foundation, the Hungarian Institute of Cardiology in Budapest, the V.A. Medical Center in Long Beach, California, and the University Hospital in Zurich, Switzerland. Each of these databases contains patient records with various medical attributes, totaling 76 features. However, most research has focused on a subset of 14 key attributes to diagnose the presence of heart disease. he dataset is relatively small, with each database containing a few hundred records. For example, the Cleveland database includes 303 instances. Given the number of attributes and instances, the dataset’s size is minimal, making it easily manageable for analysis without requiring significant storage resources. The data was collected over several years, primarily during the 1980s.

Each patient record in the dataset includes the following 14 attributes commonly used in research:​

  • Age: Age of the patient in years.​
  • Sex: Gender of the patient (1 = male; 0 = female).​
  • Chest Pain Type (cp): Categorical variable indicating the type of chest pain experienced, with values ranging from 0 to 3.​
  • Resting Blood Pressure (trestbps): Resting blood pressure in mm Hg upon hospital admission.​
  • Serum Cholesterol (chol): Serum cholesterol level in mg/dl.​
  • Fasting Blood Sugar (fbs): Binary variable indicating if fasting blood sugar is greater than 120 mg/dl (1 = true; 0 = false).​
  • Resting Electrocardiographic Results (restecg): Categorical variable with values 0 to 2 indicating ECG results.​
  • Maximum Heart Rate Achieved (thalach): Maximum heart rate achieved during exercise.​
  • Exercise-Induced Angina (exang): Binary variable indicating if exercise-induced angina occurred (1 = yes; 0 = no).​
  • ST Depression (oldpeak): ST depression induced by exercise relative to rest.​
  • Slope of the Peak Exercise ST Segment (slope): Categorical variable with values 0 to 2.​
  • Number of Major Vessels Colored by Fluoroscopy (ca): Integer value ranging from 0 to 3.​
  • Thalassemia (thal): Categorical variable indicating blood disorder status (3 = normal; 6 = fixed defect; 7 = reversible defect).​
  • Diagnosis of Heart Disease (target): Integer value ranging from 0 to 4, indicating the presence and severity of heart disease.
FilmTV movies dataset

FilmTV movies dataset

Stefan
Open Dataset
Creators: (Leone, Stefano)
Publication Date: 2018
Creators: (Leone, Stefano)

The FilmTV movies dataset serves as a valuable resource for researchers, data analysts, and movie enthusiasts interested in exploring various aspects of cinema. With data spanning over a century, the dataset provides a broad temporal view of film trends, genre popularity, and audience reception. Movies data are available on websites such as IMDb with average votes, vote numbers, reviews and descriptions. While IMDb is the most trustworthy source for data, other websites as FilmTV can provide the information on how users from different countries rate the movies compared to each other. The dataset is 0.11 GB large.

Each row represents a movie available on FilmTV.it, with the original title, year, genre, duration, country, director, actors, average vote and votes.
The file in the English version contains 37,711 movies and 19 attributes, while the Italian version contains one extra-attribute for the local title used when the movie was published in Italy.

The data set includes movies from: 1897 – 2023. Data has been scraped from the publicly available website https://www.filmtv.it as of 2023-10-21.

Food.com Recipes and Interactions

Food.com Recipes and Interactions

Stefan
Open Dataset
Creators: Shuyang, Li
Publication Date: 2019
Creators: Shuyang, Li

The Food.com Recipes and Interactions dataset is a large-scale collection of culinary data, comprising over 180,000 recipes and 700,000 user reviews spanning an 18-year period. Compiled by Shuyang Li and published in 2019, this dataset provides a rich source of information for studying user interactions, culinary trends, and recipe recommendation systems. It originates from Food.com (formerly GeniusKitchen), one of the largest online recipe-sharing platforms, making it a valuable resource for researchers and practitioners in food science, natural language processing, and user behavior analysis. The data set is 0.89 GB large and consists of two primary components: recipe data and user interaction data. The recipe data contains structured information about each recipe, including the recipe ID, name tokens (a tokenized version of the recipe title), ingredient tokens, steps tokens (instructions in tokenized form), cooking techniques, caloric level, and ingredient IDs corresponding to specific ingredients. These features allow for deep analysis of how recipes are structured, categorized, and consumed over time. The user interaction data captures engagement metrics, tracking how users interact with recipes. It includes the user ID, a list of recipes reviewed, the number of items reviewed, the ratings assigned, and the total number of ratings provided by each user. This structure enables research into user preference modeling, recipe popularity trends, and the development of personalized recommendation systems for recipe suggestions.

Rotten Tomatoes movies and critic reviews dataset

Rotten Tomatoes movies and critic reviews dataset

Stefan
Open Dataset
Creators: (Leone, Stefano)
Publication Date: 2020
Creators: (Leone, Stefano)

The Rotten Tomatoes Movies and Critic Reviews dataset is a collection of information scraped from the Rotten Tomatoes website as of October 31, 2020. It encompasses data on over 17,000 movies, including details such as movie titles, descriptions, genres, durations, directors, actors, as well as user and critic ratings. A distinctive feature of this dataset is its ability to facilitate comparisons between audience scores (ratings from regular users) and tomatometer scores (ratings from certified critics), offering valuable insights into differing perspectives on films. In the movies dataset each record represents a movie available on Rotten Tomatoes, with the URL used for the scraping, movie tile, description, genres, duration, director, actors, users’ ratings, and critics’ ratings.
In the critics dataset each record represents a critic review published on Rotten Tomatoes, with the URL used for the scraping, critic name, review publication, date, score, and content.

Rotten Tomatoes allows to compare the ratings given by regular users (audience score) and the ratings given/reviews provided by critics (tomatometer) who are certified members of various writing guilds or film critic-associations.The dataset is 0.23 GB large.

The dataset is structured into two main components:

  1. Movies Dataset: Each record represents a movie available on Rotten Tomatoes, containing fields such as:

    • rotten_tomatoes_link: The specific URL from which the movie data was scraped.
    • movie_title: The title of the movie as displayed on the Rotten Tomatoes website.
    • movie_info: A brief description of the movie.
    • genres: The genres associated with the movie, separated by commas if multiple.
    • original_release_date: The date on which the movie was originally released.
    • content_rating: The category indicating the movie’s suitability for different audiences.
    • critics_consensus: Comments from Rotten Tomatoes summarizing critics’ opinions.

  2. Critics Dataset: Each record represents a critic’s review published on Rotten Tomatoes, including details such as:

    • critic_name: The name of the critic who reviewed the movie.
    • top_critic: A boolean value indicating whether the critic is classified as a top critic.
    • publisher_name: The name of the publication for which the critic works.
    • review_type: Specifies whether the review was labeled as ‘fresh’ or ‘rotten’.
    • review_score: The score provided by the critic for the movie.
    • review_date: The date when the review was published.
    • review_content: The content of the review.

Sign In

New to BERD@NFDI? Register

Forgot password?

Register

Reset Password

Please enter your username or email address, you will receive a link to create a new password via email.