What is Learning Nodes?
How do I use Learning Nodes?
FAQs
Terms & ConditionsComing Soon
FAQReading material
Reading materialYou are sitting on a lot of data, period. That's the sole reason Learning Nodes came to be. The data marketplace is a place to connect data silos with minimal effort and bring what a free market provides best to the data world: Quality.DefinitionsThe datasetThe dataset is the basic unit of measure in a data stream, and so it is the basic unit of measure in the data supply chain.Feature vectorSince a dataset is a table with columns (features) and rows (observations), each observation is characterized by its features and is called a Feature Vector.
Dataset attributesFormatsDatasets can be stored in different formats such as .xlsx; .csv; .json; etc. and can be multiple files appended.Multiple filesA dataset will be able to append additional feature vectors as long as the data in the table being appended has the same columns (name and type).NULL valuesColumns have NULL values which indicate that the feature vector does not hold information about that particular feature. It is important to note that a NULL value is different than a zero value.To understand how a NULL (blank) value is different than a Zero (0) value let's look at an example:Alice creates a survey to gather information from her students. She asks her students for their name, age, favorite ice cream flavor. Some of the students don't want to share their age so they leave the field blank. If instead of a NULL value for students that omitted their age, we enter a Zero, then we'd be falsely classifying students in the survey as toddlers.Learning Nodes measures the NULL values present in each column to allow purchasing parties to understand the integrity of the dataset.Functionality
The data marketplace can be divided into:Datasets (Sell Orders)Datasets can be uploaded as an individual sell order or against a buy order.They are characterized by the following:IndexOut of all the columns in a dataset, the index column is the one that will hold one unique value per observation. It can be a timestamp, a date (in any format -- year/month/day; year/week; etc.), a receipt number, an event number, a unique identifier, and so on.Index typeThis is the description of the column that act as a primary key.Price per row (ppr)The ppr is the price per observation of the feature vector occurring in a dataset.Buy OrdersBuy orders are escrows that hold liquidity for datasets.The data marketplace has three main functions:Upload and sell datasetsShare your valuable data with the world by uploading it to the marketplace. Simply select a column as the index and specify a price per row.
Purchase datasetsBrowse the available datasets and purchase the quantity of rows you need. Whether you're looking for public, unlisted, or private datasets, you'll find them here.
Upload and execute buy ordersLooking to purchase datasets? Create a buy order and let data producers come to you with their offers. Choose the best dataset that fits your needs.
The escrow functionality is WIP.
