Overview

With the famous Prefontaine Classic taking place last weekend as the host of the World Athletics Diamond League final, the 2023 track and field season is coming the an end. Three world records were broken at this last world class outdoor track meet of the year [1]. This year has also seen a total of 28 world records broken on many categories (indoor/outdoor, senior/u20, men/women/mixed), in addition to memorable competitions including the World Championships held in Budapest, mid August. This leads to ask, how do these performances place on the all time rankings? How do these records compare with the recent and outstanding records set by Usain Bolt on 100 and 200 meters over ten years ago?

This dataset includes close to 500k World Athletics results of all disciplines from athletes of around 200 countries which could later be used to build models. When will Gabriele Reinsch's best all-time performance [2] will be broken or, will it ever be beaten? What are the best conditions (age, country, time of year, etc..) to optimize performance on a given discipline? Is there an ideal roadmap to make an Olympic final?

Methodology

1. Get pages references from source code of the top lists landing page
2. Scrape results for each discipline in parallel using ThreadPoolExecutor from concurrent.futures module
3. Arrange data (add some features, change dtypes, rename columns and set all-time rank)

Source: World Athletics - All time Top lists

Parameters: regionType = world, windReading = regular, bestResultsOnly = false

Data Info

Number of entries: 489k
dtypes: Int64(1), category(7), datetime64[ns](2), float64(3), int64(1), object(7)
Columns:

• all_time_rank - All-time rank based on highest results_score, needs to be greater than 1130 (Int64)
• results_score - Score from Scoring Tables of Athletics to account for equivalent performances across events (float64)
• event - Detailed event or discipline label, includes gender and, optionally if indoor (category)
• category - Detailed category label, describes the Top List from which result was collected, i.e. includes environment, age category, gender (category)
• event_rank - Rank for by Event/Category (int64)
• mark - Mark established. Units vary from time in seconds to hours or, meters or, points, according to event type (object)
• competitor - Competitor's name (object)
• nat - Competitor's nationality (object)
• date_of_birth - Competitor's date of birth (datetime64[ns])
• pos - Competitor's position in contest or race when result was recorded (object)
• date - Date when performance was achieved (datetime64[ns])
• venue - Venue where performance was achieved (object)
• age - Competitor's age at time when result was recorded (float64)
• wind - Wind information for sprints, jumps. Only performances with legal wind of +2.0 are included (float64)
• mark_details - Optional details for results converted from hand time or marks breakdown of combined events (object)

Additional Features:

• year_of_birth - Competitor's year of birth (object)
• event_name - Event short name; examples: pole-vault, 200-metres (category)
• event_type - sprints, hurdles, middle-long, jumps, throws , combined-events, race-walks, relays, road-running (category)
• environment - indoor or outdoor (category)
• age_category - senior, u20 (category)
• gender - men, women, mixed (category)

Note: Data types (dtypes) are only relevant if the parquet file is used to read data.

---

[1] World records set at the Diamond League's Finale (Sept 16-17, Hayward Field, Eugene, OR):

1. Men's Pole Vault, Armand Duplantis, Sweden, 6.23
2. Women's 5000m, Gudaf Tsegay, Ethiopia, 14:00.21
3. Men's Mile (U20), Reynold Kipkorir Cheruiyot, Kenya, 3:48.06

 Gudaf Tsegay and Armand Duplantis in Eugene, Oregon, United States of America, September 2023

 

 Current all-time ranking leader: Gabriele Reinsch in Neubrandenburg, East Germany, July 1988