Dark Mode
Spaceship Titanic ML Ready Data
Data Science and Analytics
Tags and Keywords
Trusted By
"No reviews yet"
Free
About
Data preprocessed and engineered specifically for the Spaceship Titanic prediction challenge. This resource offers instant readiness for machine learning classification tasks, featuring robust imputation of all missing values and encoding of all categorical features. Key engineered features such as CabinDeck, CabinSide, GroupSize, and TotalExpense are included, along with pre-calculated folds suitable for immediate cross-validation and model training straight out of the box.
Columns
- PassengerId: Unique identifiers for each traveller.
- CryoSleep: Indicator of whether the passenger opted for induced hibernation, converted to an integer.
- Age: The passenger's age, which has been standardised and had all missing values imputed.
- VIP: Indicator showing if the passenger held VIP status, converted to an integer.
- RoomService, FoodCourt, ShoppingMall, Spa, VRDeck: Standardised and imputed figures representing money spent in various on-ship amenities.
- TotalExpense: The aggregated money spent by the passenger across all expenditure categories, ignoring missing expenditure values.
- Alone: Indicator showing if the passenger was travelling without a group, converted to an integer.
- CabinSide: Indicates which side of the spaceship the passenger's cabin was located on.
- HomePlanet_*: One-hot encoded columns detailing the passenger's source planet (e.g., Earth, Europa, Mars).
- Destination_*: One-hot encoded columns detailing the passenger's destination planet (e.g., TRAPPIST-1e, 55 Cancri e).
- GroupSize_*: One-hot encoded columns detailing the size of the group the passenger travelled in (up to size 8).
- CabinDeck_*: One-hot encoded columns specifying the deck level of the passenger's cabin (e.g., Deck A through G, T).
- *_*missing: Several indicator columns (e.g., RoomService_missing, Cabin_missing) noting whether the original feature value was absent prior to imputation.
Distribution
The data files are typically formatted as CSV files. The sample test file includes 40 distinct features and 4277 records, totalling approximately 1.18 MB. The full offering includes four distinct pairs of prepared training and test files, some of which contain label-encoded versions of the engineered CabinNum and GroupId features.
Usage
This dataset is ideal for predictive modelling, specifically machine learning classification exercises. It is highly suitable for participants in data science challenges focused on prediction. The pre-engineered features and calculated folds make it perfect for rapid algorithm benchmarking, model evaluation, and educational study of feature transformation techniques.
Coverage
The scope covers detailed records of passengers from the fictional Spaceship Titanic. Geographic coverage includes the origin planets (Earth, Europa, Mars) and several galactic destinations. Coverage also encompasses passenger demographics (age, VIP status, group size), financial expenditure aboard the ship, and details extracted from cabin assignments.
License
CC0: Public Domain
Who Can Use It
- Data Scientists: For developing and tuning classification models using clean, pre-processed inputs.
- Machine Learning Engineers: Seeking benchmark-ready tabular data for rapid algorithm testing.
- Students and Educators: Utilising a clear example of advanced feature engineering, imputation, and encoding practices.
- Competition Participants: Requiring data structured optimally for immediate predictive analysis.
Dataset Name Suggestions
- Spaceship Titanic ML Ready Data
- Pre-processed Galactic Passenger Records
- Feature Engineered Titanic Data
- Cleaned Predictive Passenger Set
- Optimised Classification Data
Attributes
Original Data Source: Spaceship Titanic ML Ready Data
QUALITYLoading...