FIE Fencing Womens Foil Data

### Context

The International Fencing Federation (FIE) is recognized by the International Olympic Committee (IOC) as the world governing body of fencing, as such is charged with establishing the rules and implementation for international competition. It also maintains a website where it posts many international tournament results for sport fencing over the years. In addition, it has fencer bio pages for each athlete which stores biographical data.

Within a fencing tournament, there are two phases: pools and direct elimination. During pools, fencers are split into groups of 5-7 and fence a round robin (each person fences every other person in their pool) of short bouts to 5 points. The results from these pools are then tabulated to form a ranking going into the direct elimination round (DEs), where fencers face off in an elimination bracket and fencer longer matches to 15 points.

Strategies for pools and DEs vary and as such results from one cannot be directly applied to the other. However of the two, pools generates a larger amount of data (21 bouts versus 6 for 7 fencers).

You can read more about the raw data collection [here](https://github.com/amichaelsen/fie-fencing-dataset) and the initial data cleaning [here](https://github.com/amichaelsen/fencing-data-analysis/blob/main/data-cleanup.ipynb).

### Content

The data contains pool results from almost 50,000 bouts from 218 tournaments in Women's Foil from January 2014 to May 2021. Alongside this is biographical data and historical rankings data for the 2,108 fencers appearing in these bouts.

### Inspiration

Some possible questions that could be explored with this data are:
* What factors predicts wins? When one or both fencers are rated (i.e. have points), how good a predictor are these ratings? If neither fencer is rated, what are the key contributors to predicting a winner?
* How common are upsets? Are some fencer's more prone to winning upsets than losing them? Are some fencers less likely to upsets overall?
* What impact does fencing in one's home country have on performance? What about time zone differences?
* Do fencers perform 'differently' (needs to be formulated first) against fencers from the same country than competitors from other countries?
* How do fencers with rankings in multiple weapons compare with those that are only ranked in foil? What about those that are only ranked in non-foil weapons?

{'Context': '国际击剑联合会（FIE）被国际奥林匹克委员会（IOC）认定为击剑项目的世界性管理机构，因此负责制定国际比赛的规则和实施标准。联合会还维护着一个网站，该网站发布了多年来的许多国际击剑赛事结果。此外，它还为每位运动员建立了击剑者个人资料页面，存储其生平数据。

在击剑比赛中，分为两个阶段：小组赛和直接淘汰赛。在小组赛中，击剑手被分为每组5至7人的小组，进行循环赛（每人需与其组内其他击剑手各进行一场比赛），比赛时长为5点制。小组赛的结果将被汇总，以形成进入直接淘汰赛（DEs）的排名。在直接淘汰赛中，击剑手将在淘汰赛中对抗，进行更长的比赛，比赛时长为15点制。

小组赛和直接淘汰赛的战略各不相同，因此前者产生的数据量较大（21场对阵6场，针对7名击剑手）。

您可以在此处阅读有关原始数据收集的更多信息[链接](https://github.com/amichaelsen/fie-fencing-dataset)，以及初步数据清洗的详细信息[链接](https://github.com/amichaelsen/fencing-data-analysis/blob/main/data-cleanup.ipynb)。', 'Content': '该数据集包含从2014年1月至2021年5月期间，来自218场女子花剑比赛的近50,000场小组赛结果。此外，还包括2,108名参赛击剑手的生平数据和历史排名数据。', 'Inspiration': '利用这些数据，我们可以探讨以下问题：
* 哪些因素能够预测胜利？当一方或双方击剑手有评级（即拥有积分）时，这些评级如何作为预测指标？如果双方击剑手都没有评级，预测胜者的关键因素是什么？
* 意外获胜的情况有多普遍？是否有些击剑手更倾向于赢得意外胜利而非遭遇意外失败？是否有些击剑手总体上不太可能遭遇意外？
* 在本国击剑对表现有何影响？时区差异又如何？
* 击剑手在面对来自同一国家的对手时是否表现‘不同’（需要首先进行界定）？
* 拥有多个武器排名的击剑手与仅拥有花剑排名的击剑手相比如何？那些仅拥有非花剑武器排名的击剑手呢？'}