Nie znaleziono pasujących wyników.

Zalecamy wypróbowanie następujących rozwiązań, aby znaleźć to, czego szukasz:

  • Sprawdź pisownię wyszukiwania słowa kluczowego.
  • Użyj synonimów dla wpisanego słowa kluczowego, na przykład spróbuj wpisać „aplikacja” zamiast „oprogramowanie”.
  • Rozpocznij nowe wyszukiwanie.
Skontaktuj się z nami Zaloguj się do Oracle Cloud
  1. Oracle Polska
  2. Analityka biznesowa
  3. Analytics Platform
  4. Capabilities Explorer
Analytics Platform Capabilities Explorer

Data prep

Prepare and enrich capabilities help business users transform data sets from raw data gathered from source systems into data that can support analytics stories.

Topics

Preparing and enriching source data

Data from source systems is often not ready for business analytics. Business users are required to transform and enrich their data sets to create all the calculations, dimensionality, and other relevant data to support their reporting needs. This process is the last mile of data preparation that business users must undertake before starting their analytics. Users typically begin with data exports and then complete the last mile of data preparation in spreadsheets. Spreadsheets, however, are not secure and increase the potential for introducing human error. Users should be able to perform all necessary last-mile data preparation and enrichment tasks within a single secure platform that provides tracking and scheduling and removes reliance on individuals or bespoke, ungoverned spreadsheets processes.

Automated data profiling and recommendations

Whether importing data from files or connecting to existing sources, the Quality Insights capability accelerates the preparation of data for analysis. Using the power of the semantic profiler in Oracle Analytics, Quality Insights provides a visual representation of your data's quality, helping to rapidly identify issues. These indicators are based on null values, data type inconsistencies, and semantic type classifications. Each column in your data has a contextual visual to show the distribution of the represented data. Make inline edits to quickly address any issues, rename columns, and use the scrollable mini map to easily traverse long lists.

Figure 1: Data Quality Insights visual overview

Built-in data flows

Data flows provide business users with a code-free capability to transform data sets into the information needed for analytics. Connect multiple data sources into cohesive subject areas, whether your sources are in the cloud, on-premises, or personal data extracts. Enrich data through a variety of transformation capabilities, including training and executing machine learning models. Results can be saved in Oracle Analytics storage, a connected RDBMS, or Oracle Essbase. Once built and tested, data flows can be shared with other users or scheduled for regular execution.

Figure 2: Building a data flow
Figure 3: Using Regex in Data Set Editor (2:48)

Custom calculations and regular expressions

Add additional columns to data sets to create calculations. Simple drag-and-drop capabilities with Excel-like formulas allow business users to create sophisticated data sets with all the metrics needed. When more complex transformations require custom logic, Oracle Analytics allows users to experience the full potential of regular expressions (also known as Regex) through pattern matching of column text in data preparation or data set editors.

Read more about using Regex

Machine learning for data scientists

The Oracle Analytics platform has machine learning (ML) embedded into the analytics process and supports every role, from clickers to coders. For business user-friendly, no-code ML capabilities, see section Visualize.

Within data flows, choose from training numeric prediction, multi-classifier, binary classifier, or clustering with different built-in algorithms. These ML algorithms can be customized, trained, tuned, and then published to the wider analytics user community. Once models are published, they can be applied to new corporate or personal data sets.

Figure 4: Customizing a linear regression ML mode

Models trained within Oracle Analytics Cloud can be checked for quality and accuracy. For example, this naïve Bayes binary classifier has been trained with attrition data, and the quality of the algorithm is rated against the known true values from the test data.

Figure 5: Reviewing an ML model’s accuracy

Access the depth and sophistication of Oracle Machine Learning (OML), which is part of the Oracle Autonomous Database (ADW). OML Oracle Machine Learning provides a platform to leverage languages, such as R and Python, to develop, test, and publish ML models. Published models can then be registered into Oracle Analytics Cloud for wider business populations to access and execute with their own data sets. OML Oracle Machine Learning provides centrally governed models and allows gives business users the flexibility of self-service when preparing their data.

Next: Visualize