Being an avid sports buff, I will be working with data pertaining to the participants, events, winners, etc from the last 120 years of Olympics, obtained from SportsStats. Even though the dataset holds vast categories of data and can be of use to numerous people across different walks of life, I will be doing an analysis to obtain some trivia, insight into the winning trends and explain these with the data available.
The dataset used for the analysis can be obtained here
The following is the test case that I took up and worked on: I am looking to study the top performing countries and find some reason for their success while looking at other crucial aspects that determine winning. My client is any sports magazine/media who is interested in writing/running a feature on the history of Olympics and a look back at time. My target audience is sports history buffs who are into documentaries, reading and are interested in some trivia.
The data was available in the form of 2 CSV files. 1 CSV file contained data pertaining to the participating countries and regions while the other CSV file contained all the details about the events, winners, athletes, etc. I downloaded these files and uploaded them back on the Mode Analytics Platform in which I used SQL queries to wrangle the data according to the necessity.
Having set the data on Mode Analytics and cleaning it, I then did an initial Exploratory Data Analysis to uncover some prima-facie insights. To begin with, I looked at the trend of overall medals won by the top 5 performing countries. I then wanted to do a detailed trend analysis on the winning pattern, sport, gender and various other factors. I then went on to prepare a project proposal document that outlines the purpose of my analysis, an ER diagram that describes the relationship between the 2 tables mentioned above, details some of my initial observations, the hypotheses I had made, as well as the approach to vetify the same.
Each of the hypotheses were then validated, and along with it some of the initial questions that I raised during EDA were also answered. The medal winning trend on a sport, gender and seasonal level was taken and my understanding of the data was more refined. A presentation on the entire workflow of the project is found here and all the queries for each of the scenarios can be found here
There are a lot of other relationships left to explore. For example, looking at the top 10 sports of the countries, they have something to do with the popularity and cultural significance in the country. We can also do a historical analysis and the journey of a specific country, looking at their star athletes’ performances, explore some interesting incidents and stats, etc.
The dataset offers limitless possibilities for analysis for sports buffs, healthcare professionals and people interested in trivia and history. I was able to do a detailed trend analysis for United States, Great Britain, France and Germany. The US won most medals because they sent most participants. Majority of the medals are concentrated in a select few sports. There is an up-down trajectory in the medals won across history curve because of difference in seasonal games and participation. Irrespective of sex, most of the sports that contribute to top medal winning counts are similar.