Go Beyond Simple Reporting
To compete, today's organizations must rely on data-driven decision making. Unfortunately, conventional spreadsheets and databases don't know how to automatically extract valuable knowledge from raw data. For years, large companies (ex: banks, insurers, retailers) have reaped the benefits of data mining to optimize their business. However, because of tremendous complexity and high costs, automated analysis has remained out of reach for small and medium businesses. Data Applied is changing the rules of the game by offering affordable, Web-based Data Mining and Business Intelligence solutions. Did we mention it's all programmable?
|
Consider a data set containing hundreds or thousands of rows such as these: |
|
||||||||||||||||||||||||||||||||||||||||||||||||
|
|||||||||||||||||||||||||||||||||||||||||||||||||
|
Unless you're a banker, your own data probably looks quite different. However, suppose your goal is to understand which customers default on credit. Using conventional spreadsheet software, you would need to plot hundreds of charts to see how different combinations of variables affect credit default (ex: married + male + age 20-25 + two cars = 80% default). By comparison, Data Applied's automated analysis systematically examines all combinations, and finds groups with a higher probability of default. Other types of automated analysis are also possible: spot anomalies, categorize customers, pinpoint key influences, compare by demographics, etc. It's time to go beyond simple reporting! |
|||||||||||||||||||||||||||||||||||||||||||||||||
|
Other solutions: |
Data Applied: |
|
|
|
Share Data and Analyses
Perform collaborative analysis by securely sharing data sets with others. Decide who has access to what, and with which access level. Export analysis results to a searchable timeline. Tag data and analysis result with comments and keywords. Download analysis results as text or as presentation-ready images. Visualize & rate results exported to the gallery.
Visualize Large Data Sets
Visualize large amounts of data using tree maps. Rectangular tiles are used to represent individual records and obtain a global perspective of the data. Users can specify which fields should be used to color and size each tile. In addition, tiles can also be organized into groups. Search can be used to perform drill down explorations.
Identify Hidden Associations
Automatically identify associations hidden in your data. Associations which are stronger than expected (assuming independence between variables) are identified and reported. Each association is fully specified by its strength, the number of records affected, and its preconditions. Users can easily retrieve the list of matching records for further analysis.
Visualize Correlations
Automatically compute and visualize correlations between numeric fields. Correlations between each pair of numeric fields is calculated, and rendered as a graph. Filtering, area selection, and automatic group detection makes it possible to easily identify groups of highly correlated fields, or fields with little influence on others.
Discover Similarities
Automatically visualize similarity between records. Records are projected onto a two-dimensional map, in a way which attempts to project similar records to closeby cells. As a result, users can easily find which records are similar to any given one. The group identification feature allows discovery of broad categories of similar records, while the cell profile feature makes it possible to view the profile associated with each cell.
Detect Anomalies
Automatically detect true anomalies, or valid but exceptional cases. Using a powerful algorithm, records which are the most different from their closest neighbors are quickly and systematically identified. Users can visualize those records and their neighbors, for example to address data quality or fraud issues.
Build Powerful Reports
Generate pivot reports on steroid using multi-level grouping. Each intersection between a vertical and horizontal group yields a different chart. This allows users to perform visual comparisons between groups. Users can easily switch between scatter plot, bar chart, pie chart, graph line, and heat map visualizations. Search can also be used to further refine views.
Pinpoint Key Influences
Automatically determine which fields most influence others, and under which conditions. Decision trees show how the distribution of a given target variable changes, depending on values taken by other fields. For the field having the most impact on the target variable's distribution, a new split is created, showing how the target distribution was affected.
Transform Your Data
Easily prepare or transform data by specifying a set of transformation steps. Create, rename, delete, or convert fields. Filter, sample, or rank rows. Scramble values, compare values, use complex formulas and dozens of built-in functions. Use rule-based editing to execute powerful data transformation without having to write a single line of code or script.
Forecast Time Series
Automatically forecast time series. A neural network algorithm is trained using all available data, allowing it to detect and learn patterns such as seasonal variations, underlying general trends, or user activity patterns. Monte Carlo error simulations are used to evaluate the stability of the model, and visualize possible variations.
Categorize Records
Automatically discover broad categories of records using a cluster detection algorithm. The clustering algorithm finds groups of records sharing common traits, and generates profile information for each group. Users can then visualize these clusters as nodes, as well as their resemblances or differences. For each cluster, users can also view a profile of member records, and a profile of records lying outside of the cluster.
Protect Your Data
Users can be granted different access levels to different workspaces and data sets (ex: grant read, write, or delegate access). Security is enforced at the server level, so there is no way to bypass access checks. As a result, security is also enforced for any applications which uses the XML Web API to perform remote data analysis or processing.
