Exploring GRID Data with Datawisp

Authors: Mark Bate (Snr. Solutions Architect @ GRID), Cam Priest (Data Scientist @ Datawisp), Mo Hallaba (CEO @ Datawisp)

This blog post is the first in a series of posts presenting various tools and applications that can help with analyzing in-game data available from the GRID Data Platform.

Professional players will often review gameplay videos of their games to understand their performance but, increasingly, we’re seeing more and more teams utilizing in-game data to give them a clearer view of the game. In-game data refers to events (wins, losses, kills, item usage, etc.) being emitted directly from the game, in real-time.

Video reviews allow teams to understand what their players saw, but in-game data allows them to understand what the game server saw.

GRID works closely with game developers to help make in-game data available to their player communities via The GRID Data Platform. Through the GRID Portal and GRID Game Data API, GRID provides players, analysts, developers, media and others access to in-game events, current game state and statistics. Professional players and organizations who use the GRID Data Portals vary in terms of their technical abilities. That’s why GRID provides them access to the raw standardized data feeds but also to a set of tools to facilitate their esports data analysis, including the Analysis Console and Events Explorer.

The Analysis Console allows professional players and data analysts to quickly and easily review stats & events across a series.

GRID designed these tools to cover a lot of common use cases for esports teams and developers, with a specific focus on making them game-title agnostic. The tools should function the same, regardless of whether you’re viewing data from VALORANT, R6 Siege, League of Legends, PUBG or any other game title available through the GRID Data Platform.

Getting Started with Datawisp

This article highlights one such tool from our friends at Datawisp. Datawisp have built a visual platform for analyzing data from databases, data warehouses and common data formats such as Excel, CSV, and JSON. Their platform uses a visual scripting interface, allowing you to transform data by connecting the output of one block to the input of another.

Datawisp’s visual scripting interface allows users to quickly transform data from a series.

Datawisp is a paid subscription-based service, but you can head to datawisp.io to get a 14-day free trial. You can find their pricing list at: https://www.datawisp.io/pricing.

Once registered and logged in, you’ll be greeted by three options: Import data, Ask Wispy and Start analyzing. If you haven’t connected to your data yet, you can import it at this point. Otherwise, you can start using Wispy (Datawisp’s AI assistant) or start analyzing data on your own.

For the purpose of presenting how GRID Data can be analyzed using Datawisp tools, we’ll explore the series state dataset from the VALORANT Champions 2023 final (26th August 2023, Paper Rex v Evil Geniuses). Esports teams can find that dataset in the Match History Viewer by setting the date & team filters, or by following this link.

GRID Match History Viewer provides a quick overview of a series and access to the associated data files.

Once you have your dataset downloaded, head to the Data tab in Datawisp, select Import and upload the GRID post series state file you downloaded. Click the “Start Analyzing” button to create a new sheet using the imported dataset as a source.

Before we start working with the data, we should cover the structure of a GRID series state file. Let’s take a look at the dataset we imported. We confirm it’s imported as a JSON file and currently it looks like we only have one column called “value”. This implies we need to use Datawisp’s JSON handling tools to work with the data.

!

In Datawisp, you can use the JSON block to “flatten” this nested data into a more table-like format. Working with tables is often easier and we will be using this block a lot.

The Series State data is structured as a nested JSON format. At the top, you have metadata about the series itself. Stepping into the nested levels, you will also find data relating to games, segments, teams and players.

Each JSON block digs one level deeper than the last. Our single match has 4 games, so after the first JSON block flattens the games entity, mapping each game to a row, our dataset now has 4 rows.

By clicking the + button after your Data Source block or by right-clicking anywhere on the page, a menu of Datawisp’s blocks appears allowing you to string blocks together for your analysis. Here, we’ll use the JSON block to begin our flattening.

Starting with our team stats, we unpack each stat as its own column using “Automatically fetch columns.” This creates stats including team name, deaths and kills that we’ll use, along with many more options, such as headshots. From here, we’re ready to calculate K/D per team.

Conclusion

Hopefully this article has given you a good idea of how to use Datawisp’s tool and sparked some ideas that you might want to start diving into.

The Datawisp team has prepared a few templates to help you get started. Opening the links will show you how the blocks are connected and the results they got for the VALORANT series. Clicking the “Create Sheet from Template” button will clone the worksheet into your account, allowing you to run the calculations with your own datasets.

Useful Templates:

• K/D/A — calculates the kills, deaths and assists per player across the series
• Net Worth — calculates the average net worth and kills for each team across the series.