New Features

Platform & Studio

• New dataset creation feature for Twitter, PostgresSQL, SQL, MySQL, Oracle, Excel, CSV, Google News.
• Create dataset from a local TXT file using delimiter option. Supported delimiters are Semicolon, Pipe, Comma, Tab, Space
• Data Preparation: Functionality to collect, clean, manipulate raw data. User can use this functionality to make his dataset ready for analysis. Functions include –
• Aggregation
• Expression Evaluation
• Filtering
• Data Merger
• Sorting
• Descriptive statistics
• Missing Value Imputation
• Outlier Detection
• New dataset creation feature for Twitter, PostgresSQL, SQL, MySQL, Oracle, Excel, CSV, Google News.
• Train-Test split sampling
• Statistical Analysis –
  • Pearson Correlation
  • Hypothesis Test
  • One Sample Proportion Test
  • One Sample t Test
  • One Sample z Test
  • Two Independent Samples t Test
  • Two Paired Samples t Test
  • Two Sample z Test
• Statistical Analysis –
  • Rubipython
  • RubiR
• Models – Lists models published by the user. User can create models for classification and regression
• Writer – Template table and template files
• KNN: Only RMSE and MAPE values are shown in 'Result' tab.
• Linear Regression: Graphical representation of results is available in 'Result' tab.
• Association Rules: Association rule measures are available in the 'Data' tab.
• Classification: Graphical representation of results is available in 'Result' tab for all algorithms
• MLP Neural Network
• Decision Tree
• k Nearest Neighbor
• Binomial Logistic Regression
• Multinomial Logistic Regression
• Naive Bayes
• Random Forest
• Support Vector Machine
• Centroid Based Clustering: Graphical representation of results is available in 'Result' tab.
• Rubicast module has been introduced in this release. Rubicast has three different options such as Data Preparation, Data Exploration and Modeling.
• Data Preparation: It is used for accumulation (with missing interval imputation), missing value Imputation, and transformation. User can select basic and advanced parameters present in Data Preparation.
• Data Tab – It shows forecasted data along with the original data in tabular format.
• Result Tab – There are different tests such as Accumulation, Missing Value, Transformation and Differencing. It shows graph after running the tests.
• Modeling: It is used for forecasting. It consists of forecasting algorithms such as ARIMA, Auto ARIMA, Automated Exponential Smoothing, Holt Exponential Smoothing, Holt-Winters Exponential Smoothing, Random Walk and Simple Exponential Smoothing.
• ARIMA – User can input basic and advanced parameters present in ARIMA. In the Result tab, it shows the graph of forecasted data, Trained Model Parameters and Accuracy. In Data tab, it shows forecasted data with original data in tabular format.
• Auto ARIMA – User can input basic and advanced parameters present in ARIMA. In the Result tab, it shows the graph of forecasted data, Trained Model Parameters and Accuracy. In Data tab, it shows forecasted data with original data in tabular format.
• Automated Exponential Smoothing – User can input basic and advanced parameters present in ARIMA. In the Result tab, it shows the graph of forecasted data, Trained Model Parameters and Accuracy. In Data tab, it shows forecasted data with original data in tabular format.
• Holt Exponential Smoothing – User can input basic and advanced parameters present in ARIMA. In the Result tab, it shows the graph of forecasted data, Trained Model Parameters and Accuracy. In Data tab, it shows forecasted data with original data in tabular format.
• Data Exploration: It is used for exploring time series using Rubicast. User can only explore it. It is not available for run. If user clicks on Run button, it will fail. User can select basic and advanced parameters available for Data Exploration.
• Data Tab – It shows forecasted data along with the original data in tabular format.
• Result Tab – It has a graph which shows forecasted data along with the original data.
• Tests in Result Tab – There are different tests in Result Tab such as ACF, PACF, Ljung Box, DF Test and Cross Correlation. In ACF and PACF test, graphs are shown and in Ljung Box and DF Test, values are shown in the Result tab.
• Rubitext – Pre-processing –
• Case Converter
• Custom Words Remover
• Frequent Words Remover
• Lemmatizer
• Punctuation Remover
• Spelling Corrector
• Stemmer
• Advanced Entity Extraction
• Word Correlation
• Word Frequency
• Basic sentiment Analysis: Output data will be available in Data tab.
• Data Preparation, RubiText, Custom Scripts, Writer will have same functionality as implemented in studio. Custom scripts in Rubiflow is similar to Code Fusion in studio.
• Models published in Studio will be available in Rubiflow.

Rubisight (Beta)

Introducing a new module Rubisight on the Rubiscape platform. Rubisight is a Data Visualization module and can be used as part of the Rubiscape Data Science platform. Rubisight enables you to create creative dashboards and charts to help you understand your data better. Rubisight helps you with the data story-telling process empowering you to derive meaningful insights.

• Existing datasets within the workspace can be used for creating new dashboards.
• Dashboards can include various types of charts.
• Following chart types are available in the Widget Factory for creating dashboards:
• Column Chart
• Bar Chart
• Stacked Column Chart
• Stacked Bar Chart
• Donut Chart
• Pie Chart
• Sankey Chart
• Histogram Chart
• Bubble Chart
• Pareto Chart
• Area Chart
• Stacked Area Chart
• Line Chart
• Boxplot Chart
• Wordcloud Chart
• Treemap Chart
• Table
• Crosstable
• Image Chart
• Text Chart
• HTML Chart
• When creating a new dashboard, select a dataset on the left pane, the dataset columns will be classified into three categories viz. Dimensions (categorical variables), Measures (numerical variables) and Details (textual variables).
• User can select any of the available charts from the widget factory of above mentioned charts to configure a dashboard.
• By convention, the data is grouped-by Dimensions and Legends which are plotted on the x-axis and the aggregated Measures are plotted on y-axis.
• The following aggregation methods are provided: Sum, Mean, Min, Max, Count, Distinct Count.
• Filters can be applied to the data at the chart(widget) level and also at the dashboard(global) level.
• Result data can be sorted based on the columns either in ascending order or the descending order.
• Formatting options can be applied to each chart within the dashboard.
• Interactivity can be performed on the charts at the dashboard level, either in edit mode or in view mode.
• Dashboards can be exported as .pdf files