🚗 Roadmap

This page tracks features and improvements being considered for Boon. It is not a commitment — items here are ideas, in no particular order, and may change. Contributions and feedback are welcome on GitHub or Discord.

ID → name mappings

Several columns are exposed as raw numeric IDs that would be more ergonomic with a *_names() lookup, mirroring the shape of team_names() or patron_phase_names():

  • lane_names() — for the start_lane (players) and lane (objectives) columns. Values come from the CMsgLaneColor proto enum: 1=yellow, 3=green, 4=blue, 6=purple, 0=none.

  • hitgroup_names() — for the hitgroup_id column in damage. Source 2 uses a standard set (head, chest, stomach, limbs, …); confirming the exact set Deadlock uses needs a pass over the game data.

  • More as new opaque enums are surfaced.

Visualization

Helpers to turn parsed DataFrames into visuals without writing ad-hoc plotting code each time:

  • Static plots for common views: net-worth-over-time, kill timelines, damage-dealt-vs-taken matrices, lane control.

  • Position heatmaps per hero or per team, computed from player_ticks.

  • Animated GIFs of player movement over a match (or a window), rendered on a stylized map.

  • A small style API so plots/GIFs use consistent Deadlock colours (team, hero, lane) by default.

Analysis helpers

Composable, derived stats rather than only raw rows:

  • A match_summary helper returning per-player KDA, GPM, XPM, last-hits, hero damage, objective damage, etc. in a single frame.

  • Combat-encounter detection — grouping kills and damage into fights and tagging participants, location, and outcome.

  • Win-probability over time — an interface for plugging in a model and exposing a win_prob column on world_ticks.

Performance and ergonomics

  • Streaming / incremental parsing for partial or in-progress demos.

  • Bulk-match utilities for analysing many demos at once (parallel parse).

Have an idea?

Open a GitHub issue or chime in on the Discord.