Following ingestion and consolidation of a submitted dataset, Inspirient’s engine automatically calculates descriptive statistics and applies a set of analytical methods. In detail, these analytical methods and their input requirements are listed below.
| Analytical method | Description | Data requirements | Patterns | Search tags |
|---|---|---|---|---|
| Descriptive statistics | Calculating descriptive statistics for all dimensions of the input data and generates a profile for each column and table | Categorical and/or numeric variables | Maximum, minimum, average, standard deviation, range of values, and type of frequency distribution | Column profile Table profile Histogram |
| Aggregation | Aggregation analysis, comparable to a ‘pivot’ analysis in Microsoft Excel, for all dimension combinations (up to triple-wise dimension combinations) | Categorical and/or numeric variables | Maximum, minimum, average, and standard deviation | Aggregation |
| Anomaly detection | Discovery of potentially unknown patterns through unsupervised machine learning techniques for outlier detection | Business irregularities, data inconsistencies, deviations from a pattern, and column-level outliers | Categorical and/or numeric variables, business KPIs (optional) | Anomaly |
| Slicing and dicing | Automatic slicing and dicing of the dataset where appropriate, for example focussing on the last full business year or drilling down on the highest selling product | Categorical and/or numeric variables | n/a | Drill-down |
| Time-series trends and forecasting | The trend analysis provides a comprehensive ranking of all time-series trends and forecasts in the input data | Time dimension, and categorical and/or numeric variables | Time-series trends and deviations from trends | Time series Trend Deviation |
| Single- / multivariate regression analysis | The regression analysis detects relationships between variables (linear and logistic) and summarizes the most significant relations. The output table includes relationship strengths and functions | Multiple numeric variables and categorical variables | Correlation strength and anomalies | Regression Correlation Deviation |
| Geo Analytics | Analysis of geographical information and visualization as regional maps and/or geo-location highlights | Location variable | Regional hotspots | Geographic |
| Root Cause Analysis | Explanations of patterns in the input data, possibly leading to detection of hidden causalities | Dependent and independent variables | Association rules explaining relevant patters | Root cause |