The loaded data file is called a dataset. Under "File" menu, click on "Open" to load a dataset and select the dataset to load into EventFlow. Once the dataset is loaded, the total number of records is shown at the top of the Control Panel. EventFlow will also automatically load the associated configuration (config) and modification (mod) files if they exist.

Note: When a dataset is loaded and the user would like to load a different config and mod file, the user must load the config file FIRST and the mod file SECOND.

If the dataset is too large to fit into the JVM memory (memory used to execute the system), the loading process will stop and a dialog will appear to provide users with options to filter the data in the following ways:

1. Panel 1: Limit the event categories by unchecking the boxes besides the categories.
2. Panel 2: Limit the events by setting a time window; Limit the events by setting a event count.

Note: If the user set any filters in Panel 2, the user can also set the alignment point around which the filter works. For example, No Alignment means the pivotal point around which the filter works is the very first event in the record.

Similarly, if the overview building process runs out of memory, a message will appear that tells the users to close and reopen EventFlow.

Tip: Check out this EventFlow video demonstrating how to handle large and complex data: Handling Large and Complex Data.

If your analysis does not need the EventFlow overview, you can click "Open (no overview mode)" under the "File" menu to load the dataset with only the timeline view. Hiding the overview will increase EventFlow's rendering speed. All functionalities work as usual, e.g., advanced search, interval merge, and combine categories.

Tip: To bring back the overview, you need reload the dataset by clicking "Open" under the "File" menu. Remember to save your changes before reloading.

When data errors are detected (e.g. an interval end time is earlier than the interval start time) a dialog box appears.

"Save" will save the config file (with current colors choices, event category selections, options, etc., [see section on config]) and mod file (with all insertion, aggregations, replacements, etc., [see section on mod]) but does NOT touch the data file. This is useful to save your ongoing work. *NOTE: Older config and mod files will be overwritten (so make copies before saving if needed). If a config or mod file do not exist, they will be created if applicable.

"Save as" will save a new data file with only the records that are visible in the overview AFTER applying the current modifications to the data (i.e. whatever aggregation, insertion, etc. have been done will be applied - and cannot be undone). A control panel also gives the option to only save the event categories that are selected in the legend (on the overview side). This helps further reduce the amount of data saved. Users are then asked to pick a name for the new data file, and the new config file and attribute file will use the same name - with the appropriate extensions. "Save as" is very useful to create smaller datasets with only the data of interest, or apply permanently apply modifications to a dataset.

Because the modifications are applied to the new dataset (and the list of modifications is lost), you may want to use "Save" first before doing the "Save as" so you still have the modification file for those data modifications. You can also save the modification file separately (and give it a meaningful name) using "Save mod", so you can reuse it later or on another dataset.

We highly recommend that you take notes about all the modifications you do and how you filtered the data. At this point the EventFlow prototype does not have a logging feature or complete history mechanism, but we recognize that any product based on EventFlow should include such functionality

A configuration (config) file saves changes made to event categories and display options.

The following changes to event categories can be saved in a config file:

• Select event categories to display in the Overview and Timeline panels in the Legend Panel
• Color of event markers
• Sort event categories using "Sort Legend by..." in the View menu
• Order event categories by dragging and dropping in the Legend Panel

The following display options can be saved in a config file:

• Overview Options in the View menu: Episode beginning category, representative time, merge repeated events, fill gap with previous event's color, show vertical gap between sequences, and layout parameters
• Timeline Options in the View menu: Category label, and layout parameters

The following features are NOT saved:

• Name of event categories
• Alignment by: K-th Occurrence of an Event Category, Exact Date and Time, and All Occurrences of an Event Category
• Temporal Windowing in the Overview panel: Set Window by Tme and Set Window by Event Count
• Simple Search Methods: Event Count, Subsequence, and Overlap
• The rank in which records are organized in the Overview and Timeline panels

When EventFlow loads a data file, for example sample.txt, it will look for a config file with the name sample_config.xml. If that file does not exist, users can manually load other config files.

The config file is in XML format. Although not recommended because of possible syntax mistakes, users can also modify it manually outside of EventFlow.

Users can "Save config" or "Save config as" which may be useful to create different starting points (e.g. for demonstration) or to reuse the same config file for variations of the data file (e.g. an updated or larger version of the dataset as long as it contains the same event types)

Data modifications/transformations can be saved in a modification (mod) file. This is convenient when users want to perform the same modification on a similar dataset later. It will save the following:

• Rename an event category
• Category Aggregation: Simple and Advanced
• Category-Attribute Switch
• Gap and Overlap Removal (Merge) of Interval-based event categories
• Insert event at Alignment Point
• Advanced Search and Replace

When EventFlow loads a data file, for example sample.txt, it will look for a mod file called sample_mod.xml. If the mod file does not exist, no mod file will be loaded automatically. Users can manually load other mod files as long as it pertains to the same or similar dataset.

The mod file is in XML format. Although not recommended because of possible syntax mistakes, users can also modify it outside of EventFlow.

Users can "Save modifications" or "Save modification as." "Save modification as" is useful when users would like to apply different data modifications to the dataset (for demonstration).