Visual Pattern Recognition

Data Visualization, Experiment


When browsing the web we leave a lot of traces. Most servers have a log file which lists the IP-addresses of the visitors, the port they are using, the file they are requesting and the time. For system administrators these information are highly valuable to detect malicious behaviour. In the course Visual Pattern Recognition we interviewed various people dealing with these log-files on a daily basis and found out that they all used text-editors to inspect these information. Every request on a web-server produces one line in the log file – when browsing a standard website this can easily add up to dozen requests per user. Now imaging how big these text-files are and how hard it is to find valuable information in them.

To make their lives easier we created various data visualisations which made these information more accessible. As the sys-admins we worked with confirmed it was much easier to spot strange behaviour, which could e.g. be a hacking attempt, using our interactive visualisations than simply looking at the text-log.