LifeFlow is a tool designed to analyze temporal event sequences by visualization. It is recommended that the users read the paper first to become familiar about the design of LifeFlow. The paper can be found at: http://www.cs.umd.edu/localphp/hcil/tech-reports-search.php?number=2010-22
LifeFlow can be used on systems running either Mac OS X or Microsoft Windows. The minimum system requirements are:
LifeFlow has been tested on Mac OS X 10.5-10.7, Microsoft Windows XP and Windows 7.
LifeFlow is now available. Please email Catherine Plaisant at ude.dmu.sc@tnasialp for the executable and source code.
After extracting the zip file, run LifeFlow by:
Four different files are used in LifeFlow.
It is a 4-column tab-delimited text file.
Use ; to separate between attributes. Leave this column blank if you do not have any event attribute. ; ; ...
|Record ID||Event type||Date and Time||Event attributes|
|4318||Arrival||2010-01-19 03:53:00||Account No.="24287939"; Entrance-Diagnosis="Heart attack"; Exit-Diagnosis="Heart attack"|
|4318||ICU||2010-01-19 03:53:00||Room No.="125"|
|4318||Floor||2010-01-19 03:53:00||Room No.="854"|
|4318||Exit-DISC||2010-01-19 06:44:00||ED-Dispo="DISC"; Exit-Diagnosis="Heartburn"|
|5397||Arrival||2010-01-19 03:53:00||Account No.="24287939"; Entrance-Diagnosis="Heart attack"; Exit-Diagnosis="Heart attack"|
|5397||Floor||2010-01-19 03:53:00||Room No.="854"|
|5397||Exit-DISC||2010-01-19 06:44:00||ED-Dispo="DISC"; Exit-Diagnosis="Heartburn"|
This file store the non-temporal attributes of each record. You are not required have to have this file if you do not have any record attribute. It is a 3-column tab-delimited text file, which has the same name as the event data file but with extension *.attrib.
|Record ID||Attribute name||Attribute value|
The config file helps the users to bring back their previously used setting and continue the analysis (Section 4.1). For example, their preferred color of the event. Another important aspect is to carefully maintain a log file about where the data came from, what modifications were made, etc. To help convert data to the LifeFlow format we are developing the DataKitchen application.
Throughout the manual, a sample dataset is used to demonstrate the features in LifeFlow. Please email ude.dmu.sc@wtsirk for the sample file.
The sample data is a generated set of electronic medical records with six event types: