Analyzing Former vs Current BoomID Carry Players: A Data-Driven Approach

personalProject | March 11, 2023, 11:49 p.m.

Boom Esport made a significant change in the early DPC 2023 which one of the changes being Jackky the carry player leaving the team than join Bleed Esport and Natsumi came as the replacement.

This move has sparked a debate among fans, with some believing that it is a considerable downgrade while others argue that Natsumi is a competent replacement. Now that the first tour has been completed, let's examine the data to better understand both player.

Let's gather all the data

I collected all the matches that they both have done on the recent DPC tour from opendota.

Convert it to Panda Dataframe to simplify the analysis process

Preprocessing

Dropping unnecessary columns with difference function since the dataset contained a lot of columns.

To make analysis easier, we can rename the column that have long name such as 'gold_per_minute' to 'gpm', rescale 'hero_damage', 'tower_damage', and 'kda' using the min-max method to handle the wide range of values, and convert the 'result' column datatype to string and replace the values with 'win' or 'lose'.

Grouping the dataframe by result and name to make a comparison.

Data Visualization

Comparing some essential carry player skill data with bar chart.

The Dataviz above vividly portrays the distinctive styles employed by each carry player and how they differ in their approaches to both winning and losing games.

Let's try to put the universal efficiency formula to the test

For this analysis, I utilized a general efficiency equation and analyzed the dataset provided. Using the information available, I calculated two efficiency ratios, one based on Gold per Minute (GPM) and another based on duration. The resulting formula, which shows the efficiency as the ratio of input divided by GPM or duration, is presented in the image below for easy reference.

And here is how the results look like.

Conclusion and making it into radar chart
After conducting a thorough data analysis and visualizations, the results indicate that Natsumi's performance as a carry player is closely tied to whether his team is winning or losing. When his team is losing, Natsumi's teamfight participation, GPM, and last hit per minute are all lower than Jackky's, and his duration efficiency is also lower. However, Natsumi's higher GPM efficiency suggests that he takes fewer risks while farming.
In contrast, when his team is winning, Natsumi's teamfight participation, GPM, and duration efficiency are all higher than Jackky's, indicating that he is able to make meaningful contributions to his team's success. While Natsumi's farming efficiency may be slightly lower than Jackky's, his ability to contribute to teamfight and duration efficiency metrics in winning games suggests that he is still a valuable player on the team.
Overall, our data-driven approach sheds some light on the performance of Natsumi and Jackky as carry players for Boom Esport. It suggests that while Natsumi may not farm as efficiently as Jackky, his contributions in other key areas make him a valuable player.
However, it's important to note that evaluating a carry player's effectiveness is a complex process. There are many other factors that can influence a player's performance, such as objective securing, ward placement, and successful ganks. Additionally, different players and teams may have unique playstyles and strategies that can impact how efficiency is measured.
As the author, I want to emphasize that I am not a professional in Dota 2 or data analytics. While I have made every effort to ensure the accuracy of my analysis and conclusions, this article is meant to be a starting point for further exploration and discussion.

Full notebook link