SnapShots class

class zilean.SnapShots.SnapShots(timelines, frames=[8], creep_score=True, porportion=True, verbose=False)

Bases: object

SnapShots is used for extracting interesting player data from Riot MatchTimelineDto s. SnapShots is a helper object that facilitates data analysis on League of Legends matches.

The reason for the name is because SnapShots can extract player data from a match at specific time intervals, or frames (in minutes). Data at frames of interest can be used to, for example, predict the result of a match.

timelines
The source data. It can be either a:

  • String, a file name where either a list of MatchTimelineDto s (in JSON format) or the computed summary statistics (csv) is stored. The computed summary statistics (csv) should be an earlier saved DataFrame using the SnapShots.to_disk() method.

  • List. A list of MatchTimelineDto s.

  • Dict. A single MatchTimelineDto.

Type

str | list | dict

frames

Integers indicating the frames (in minutes) of interest. Default [8]. This argument does nothing if the specified input timelines file is a stored summary file in csv.

Type

list

creep_score

Compute the creep score for the players, then drop the minionKilled and jungleMinionKilled feature of the players. Defaults to True.

Type

bool

porportion

Add goldPorportion and xpPorportion as features to the players. Defaults to True.

Type

bool

verbose

Print out the progress of loading the source data, defaults to False.

Type

bool

get_lanes(lanes, per_frame=None) → list

Return a slice of the summery statistics that represents specific lanes in the game. Statistics of a specific lane in the summary is marked by an underscore and a number at the end of each feature. For example, totalGold_0 represents the total gold difference of the TOP lane.

Parameters

  • lane (list) – Position options are any of {“TOP”, “JUG”, “MID”, “BOT”, “SUP”} or their corresponding index {0, 1, 2, 3, 4}.

  • per_frame (bool, optional) – If False (default), each match (MatchTimelineDto) is one dictionary. If True, each frame (in minutes) of a match is one dictionary. Defaults to False.

Returns

A list of dictionaries including the statistics for the lanes of interest, the matchid, and the label (win).

Return type

list

summary(per_frame=False) → list

Return the summary for all the matches (MatchTimelineDto). For each match, summary statistics of every time frame of interest is returned. The summary is ready for further data analysis.

Parameters

per_frame (bool) – If False (default), each match (MatchTimelineDto) is one dictionary. If True, each frame (in minutes) of a match is one dictionary. Defaults to False.

Returns

A list of dictionaries, ready for further data analysis. Each dictionary is either a match or a frame (see per_frame).

Return type

list

to_disk(path='data/', verbose=True) → None

Save the summaries to disk as csv files using pandas.DataFrame.to_csv()

Parameters

  • path (str) – Path name relative to your working directory. Defaults to data/

  • verbose (bool) – Print the directory where of the saved file. Defaults to True