Hey Folks, Wassup!

Recently when watching a professional Dota 2 competition I saw all kinds of interesting analytics of the gameplay being shown on the screen, just like in the real sports broadcasts. That got me pretty intrigued how exactly are they extracting information from the LIVE gameplay. Are they doing this manually, or is the process automated?

Affiliate Disclaimer: A few of the resources stated in this article contain affiliate links. That means if you find these resources helpful, worthy of spending your money on and you buy a subscription, a course or a book; I get a small cut without you paying anything extra.

I recommend these resources to you because I think the content they offer is pretty good & these will assist you big time in upskilling yourself enabling you to soar in your career.

Dota data analytics

Image credit: Microsoft

I started my research, & came across the process of applying data analytics in ESports. AI-assisted data analytics is applied on raw gameplay stream, on both LIVE & historical data to make the stream more engaging for the viewers & for some other important use cases which I am going to discuss in this write-up.

Data analytics in ESports is kind of an intersection between the software development and the gaming universe. A sort-of middle ground. For those who are into software programming and want to transition into game programming. This can be an interesting field to look into. Also, it could be an exciting field for anyone interested in the data analytics domain.

In this write-up, I discuss the rise in the popularity of the application of data analytics in the domain of ESports. How lucrative is fetching analytical insights from the game streams? I also talk about the future prospects of the segment, market growth, jobs, startup opportunities etc.

I continue the discussion with the design requirements and the technical aspects of a tool called Echo implemented by Digital Creativity Labs, University of York, UK, & deployed in an international Dota competition.

So, without any further ado. Let’s get on with it.

1. ESports – Growth from A Niche Segment to A Mainstream Mode of Entertainment

The popularity graph of MMO Massive Multiplayer Online games like Dota2, Fortnite, FIFA, Clash of Clans, PUBG etc. has rocketed in recent years. These games have been super successful in attracting millions, if not billions, of gamers competing against each other either individually or in teams. Fortnite alone has over 250 million registered users & holds a record of 10.8 million concurrent users.

Tech giants have more than noticed the exponential growth in the market size of online gaming & are investing some serious money in building game streaming platforms like Mixer owned by Microsoft, YouTube Gaming owned by Google & Twitch owned by Amazon.

Twitch is the leading gaming platform with 2.2 million daily & 41K concurrent broadcasters, 15 million daily & 1.1 million concurrent viewers on average.

Both casual & professional gamers broadcast their gameplay streams LIVE on platforms like these round the clock. Besides this, ESports competitions are held across the globe every year. The 2019 International Dota 2 Championship had a prize pool of 34 million dollars & the final was viewed by over 5 million concurrent viewers.

According to a report by Newzoo, the ESports market will cross 1 billion $ mark this year with the maximum revenue coming from the brand sponsorships.

Some more statistics

According to the Nextweb, the present ESports audience is double the size of the global audience for F1 racing, eight times bigger than the TV audience for the baseball world series & ten times bigger than the number of people that watched 2019 Super Bowl.

Naturally, any segment of the market which such an exponential rise is going to attract investment from and the involvement of the big guns. And this will eventually cascade into new job opportunities, higher pay scale and other startup opportunities.

2. How Data Analytics & Mining Is Leveraged to Make ESports More Engaging & Entertaining?

Data analytics in the software industry is no more a shiny new thing. It’s key to the success of the businesses in the present times. Running analytics on the data helps us gain a better insight into the market, user behaviour, customer demands & stuff. Analytics helps us evolve our product, serve our customers better.

Application of data analytics is prevalent in the other segments of the industry like E-commerce, social networking platforms etc but is something new in the segment of ESports.

Below are the few use cases where in ESports both the real-time and the historical detailed gameplay data is leveraged to extract meaningful information.

2.1 In Making the Live Stream More Engaging

ESports streaming providers are moving towards more data-driven content production. Quick & Easy to comprehend graphical representation of the analytics data increases the user engagement on the platforms.

Analytics data is ideally statistics, timings of the player’s reaction to the in-game events & other information not visible to the naked eye. Special tools are designed that extract raw data and convert it into a graphical form easily consumable by the viewers thus enhancing the viewing experience.

Many platforms enable direct streaming of raw gameplay data stream to the user’s PC, assembling the stream into an interactive view of the match. Enabling the user to change the camera angles, go back in time, interact with the virtual objects in the game etc.

If you found the content helpful, check out the Zero to Software Architect learning track, a series of three courses I have written intending to educate you, step by step, on the domain of software architecture and distributed system design. The learning track takes you right from having no knowledge in it to making you a pro in designing large-scale distributed systems like YouTube, Netflix, Hotstar, and more.

2.2 Understanding the Competitor’s Gameplay Tactics & Skill Building

ESports is all about competition and skill-building & data analytics helps a great deal in understanding the intricacies of our gameplay and that of our competitors. Platforms & tools built for ESports enable us to record gameplays, watch them over and over, compare multiple gameplays together etc.

With analytics, we can figure out the tactics & strategies the opposite team uses often & how we can counter those. The extracted information via analytics helps us easily analyse our performance, strengths, weaknesses & enables us to have an edge over our competitors.

It’s a clear example of how analytics is directly linked to the success of a team. Also, this data helps viewers learn, as they watch a match.

Microsoft partnered up with Cloud 9 a professional esports organization based out of LA, California to help them strategize better, train with the insights obtained from the data analysis.

SAP teamed up with Team Liquid, which is a multi-regional professional esports org. based in the Netherlands to improve the player and the team performance via AI-based data analytics powered by their product SAP Hana.

SAP Hana is an in-memory, column-oriented, relational database built for both OLAP & OLTP purposes. It facilitates advanced analytics such as predictive analytics, spatial data processing, text analytics, streaming analytics, graph data processing etc. It will provide the core technology for the development of the cloud-based analytics platform. Besides this SAP will also leverage SAP Leonardo digital innovation system, IoT, predictive and machine learning techniques to help the team train better.

Intel extended its partnership up with ESL Electronic Sports League worth over 100 million dollars, delivering end to end gaming solutions.

Mobalytics is another platform which assists gamers to improve their skill in the game League of legends based on the analytics results. There are analytics platforms dedicated to specific games like Shadow, a counter-strike analytics platform, Fortbuff a Fortnite analytics platform, DotaBuff a Dota analytics platform.

2.3 Monetize Streams Better

As I stated earlier that the majority of the revenue in the ESport segment is from the brand sponsorships. Twitch has partnered up with a company named MVPindex to leverage AI-driven data analytics to monetize the game streams better. With real-time reporting and stream performance data, sponsors have better insights into the value of the content created by gamers.

Twitch uses AI & Speech recognition tech to monitor chats, video frames & audio of streams to get the accurate engagement and stream impact on the viewers. With all this information available sponsors can make an informed decision on where to put their money on.

2.4 Betting Industry

Many of the countries have legalized betting on esports with some regulations applied. Data analytics empowers the gamblers to study the trends over time. Provide people with better insights when & where to place their bets. Betting markets are powered by the analytics information extracted from the esports data.

3. Echo – An Analytics Tool to Identify Important Data Points in a Dota2 Live Stream

This is a gist of the paper published by Digital Creativity Labs, University of York, UK about a tool called Echo that uses volumes of data both LIVE & historic to detect key points from a Dota 2 game stream. It then converts the important data points to graphics easily consumable by the audience.

The tool ingests information from several sources such as the LIVE video, chat messages, feedback of surveyed audience members etc. to churn out interesting graphics making the stream and the commentary of the match more engaging.

Echo was deployed at the ESL One Hamburg 2017, one the largest esports global tournaments, watched by 25 million people online. More than 40 hours of tournament video footage was analyzed including 9 million chat messages & feedback by 98 audience members. The graphics created by the tools caused elaborate discussions amongst the commentators & improved the viewer engagement by a significant amount.


3.1 Design Requirements

Some of the key design requirements of the tool included Simplicity, the extracted information presented before the viewers should be simple enough to be quickly and easily consumed by them.

The entire data ingestion and processing process had to be automated. The processing algorithms had to be smart enough to identify interesting patterns and anomalies. Resources handling the tool should be able to easily run the whole process right from the data ingestion to the end generated graphics with just the click of the buttons.

3.2 Technology

Echo was written in C# and WPF, WPF stands for Windows Presentation Framework, it’s a GUI framework used with the .NET framework. The database used was SQLite for persisting match data. Match data was obtained from the official Dota 2 API in conjunction with some other Dota APIs. The data was processed with a customized version of Clarity, an open-source parser written in Java for Dota2.

DataCamp helps you understand the language of the data by training you on data engineering, data science & machine learning. With their skill-based tracks, you can become a Data Engineer with Python, Machine Learning Scientist with Python, R & more. The platform has two learning tracks you can either start from zero or you can advance your data science skillset if you already have knowledge on the domain. Check it out.

4. Jobs In the Esports Segment

With the increasing popularity of the data analytics, naturally, there would be job openings for that in the industry. I looked for some of the Data Engineer Job openings in the esports domain. The common requirements I found in them are –

The candidate should have a degree in computer science, applied mathematics or Information Technology;
Should have experience with analyzing large data sets, Statistics, Probability, Machine Learning.
Knowledge in SQL, Python or any other related programming language.
Experience with data handling & streaming frameworks like Hadoop, Apache Spark, Kafka, Kinesis etc.
Experience in media, esports, the entertainment industry is a plus.

The requirements are the same as the data analyst requirements in the software development industry. SQL is the most popular tech to interact with the data, one should know a backend language like Python, Java to implement the business logic. Python is more popular with machine learning engineers. Large scale streaming data has to be stored in Hadoop or somewhere. Facebook uses Hadoop to manage the crazy amount of data generated daily on the platform. Streaming data is managed by routing it through data pipelines with the help of tools like Spark, Kafka, Kinesis, Storm etc.

If you want to read more about the jobs, here they are

Senior data analyst at Riot Games, Data analyst at Cloud9

Data engineer at Esports One,  Data Analyst at Twitch

* The openings could be closed, I’ve just added the links for an insight into the job requirements.

5. Starting Our Own thing – Startup Opportunities

Besides working for a tech company, there are a lot of opportunities to start our own thing in the esports domain. The segment is relatively new, there aren’t many web-based platforms catering to the esports fans. Not that I know of.

Generally, the gamers hang out in the discord chat groups. I am not aware of a platform that connects casual gamers, enables them to challenge each other across the game streaming platforms like Twitch, YouTube etc. Notify their followers of the upcoming events. Have a global cross-platform leaderboard, provides gameplay analytics, help them promote their content. This is one pain point I am aware of. There are other manifold opportunities in esports.

Writing a fantasy-based esports app with betting integrated is a good idea. Sports-based fantasy apps & games are pretty popular with sports fans.

You can get access to the gaming data via portals like OpenDota. It’s an open-source platform providing Dota 2 data via an API. Developers can leverage the API to implement their ideas. Similar to this, there are portals available for every specific popular MMO game.

For Dota2 there are some other platforms like DotaBuff, DatDota etc. These APIs provide access to match data aggregated from tens of thousands of matches. On this data, we can run algorithms to identify successful gameplay strategies over time.

Well, Guys, this pretty much sums up the discussion on data analytics in esports. If you liked the write-up you can share it with your folks. Follow 8bitmen on social media.
You can read about me here.

If you liked the article, share it on the web. You can follow scaleyourapp.com on social media, links below, to stay notified of the new content published. I am Shivang, you can read about me here!

I’ll see you in the next write-up!!
Until then… Cheers!!

More On the Blog

Facebook Real-time Chat Architecture Scaling With Over Multi-Billion Messages Daily

Twitter’s migration to Google Cloud – An Architectural Insight

Instagram Architecture – How Does It Store & Search Billions of Images

How Does PayPal Process Billions of Messages Per Day with Reactive Streams?

How Hotstar scaled with 10.3 million concurrent users – An architectural insight