Background

Player journey maps visualize all player actions in chronological order. They can be helpful to find the sequential relationships between granular player actions. However, when incorporating increased numbers of these fundamental actions, the visualization rapidly becomes unreadable. With clutter and outliers, it is difficult to find insight from users.

INSPECT is a tool that allows users to interact with player journey maps. There are three main features that help users simplify the graph. The first interaction provides frequency-based filtering, which can remove unpopular actions and transitions. The second and the third are segmentations based on player-specific variables and player behaviors. These segmentations are designed to help separate the graph into subgraphs and help find differences to answer research questions.

Approach

INSPECT is a web application coded in Python. We use plotly Dash¹ as the main package to build the frontend and the backend. Dash allows people to build data-related analysis applications and deploy these on a web server. To support different features, we also use several other packages such as Json, pandas, datetime, and statistics, which are used for handling input files, processing data, handling temporal information, and calculating means and/or medians, respectively.

There are three interactions that are supported in INSPECT. The first interaction is to inspect the journey map. Users can click on nodes and edges, as well as filter out unpopular nodes and edges based on their settings. The second interaction is to segment players based on player-specific variables such as demographic data. The third interaction is to segment players based on their behaviors.

Visualizations

Video

Teams