Search stories are the main data object in LOIS.
All starts with the navigation log collected by the LOIS browser extension for each user solving an online information search task. Each click becomes an action in the story.
This log is then cleaned by technical biases (like automatic reloads and system events), and eventually each action is enriched by automatic tags based on domain, query, duration, etc.
Search actions are divided in search episodes, that is, portions of the story starting with a search action (on a search engine) or a system action.
Stories are then visualized in many fashions to help reesearchers see meaningful patterns, and also to make them rich instructional examples.
This user did one search (the dark blue square) and then visited more websited, jumpiing from one tto the other (the brown squares). After a while, he got back to an already visitted page (the light yellow square). Then he revised the search query (pink square) and visited other websites. We call stories with long episodes (thas is, lines with more than 3 squares), intensive stories.
This other user, working on the same task, tried out two different queries (the first two dark blue squared) visiting juust one website each time. Then he found a query he liked, and opened 7 different web pages from the same result page (all the lines starting with the light blue square). We call stories with many lines (=episodes), extensive stories.
Here we have still another search stile: this user tried different search queries, and each time goest to a website where he visits more pages (brown square followed by yellow squares).
LOIS 