This is not an official Google product.

Crowdsourced Realtime Gaming

We decided to get a little crazy and build a real time, crowdsourced, and live streamed gaming platform.

Imagine players all over the world shouting "up! down! left! right!" into their computers and phones to control a common character while chatting in different languages.

Inspired by Twitch Plays Pokémon, this game uses Firebase to crowdsource input from thousands of players in real time, then a Node.js process running on a Google Compute Engine Windows VM sends these crowdsourced commands to a game, and live streams it all to YouTube.

The game's web front end uses the power of Google Cloud Machine Learning to enable realtime voice transcription and translation.

Finally, BigQuery gives us the power of big data to analyse all the commands and find interesting patterns.

This is originally presented at Google I/O 2016. You can watch the talk here:

Setup

There are three components of this demo that run in the cloud:

• Firebase handles all of the real-time data syncing between multiple players and the game server. Firebase will handle relaying commands from every player to the server, as well as syncing chat messages between players.
• The game server is a Windows virtual machine running on Google Compute Engine. It runs the game natively, relays user commands to the game, and streams everything to YouTube Live.
• BigQuery is the cloud data analytics platform that will receive all game events, and supports using SQL to query various reports.

Set up your Firebase and Google Cloud Platform project

1. Create a new Firebase project.
2. Click Add Firebase to your web app to get the Firebase and Google Cloud Platform connection settings:
   • API key
   • Database URL
   • Project ID.

Follow the instructions here to extract your project ID.

The settings should look something like:

var config = {
  apiKey: '...',
  databaseURL: '...',
  projectId: '...'
};

Replace the placeholders in frontend/config.js with these values.

Firebase security settings

We require users to be authenticated to write messages to the chat node or read from the commands queue. These rules will be automatically deployed as they are included in the firebase.json file.

{
  "rules": {
        "chat": {
          ".read": true,
          ".write": "auth != null"
        },
        "commands-archive": {
          ".read": false,
          ".write": "auth != null"
        },
        "commands": {
        	".read": "auth != null",
        	".write": true
        }
    }
}

Firebase service account

Create a service account that you'll upload to your game server created in the next step.

1. Create a service account for your Firebase project.
2. Select Furnish a new private key and select type JSON. Rename this file to firebase-plays.json. Save this file; you'll upload it to your game server in the next section.

Enable Speech and Translate APIs

• Click here to make sure the Cloud Speech API is enabled for your project.
• Click here to make sure the Cloud Translate API is enabled for your project.

Create the game server

Create a Windows virtual machine using the Cloud Launcher. Click Launch on Compute Engine and then click Create on the next page.

Install the following on the Windows virtual machine:

• Google Chrome
• gcloud - the Cloud SDK for Google Cloud Platform.
• Node.js
• Python
• Visual Studio Express 2013 to address this MSBUILD issue.
• Install git

Optional: YouTube live streaming

1. Set up YouTube live streaming on your YouTube channel, and install an encoder on your Windows virtual machine.
2. Get your YouTube live stream channel URL and paste it into frontend/index.html.

Set up BigQuery

1. Go to the BigQuery page for your project.
2. Create a new dataset named firebase_plays.
3. Create a new table named commands. Under schema, click Edit as Text and paste in the contents of the table-schema.json file.

Deploy the web app

Make sure you have the Firebase SDK installed. You can install it with this command:

npm install -g firebase-tools

To deploy, run the following command in the frontend folder:

firebase deploy --project <YOUR-PROJECT-ID>

Run the game and Node.js game-controller app

1. Install and run whatever game you'd like to run on your game server. The script only supports the arrow keys on the keyboard by default.
2. On the Windows game server, run the following commands in the git bash terminal:
   1. git clone https://github.com/bretmcg/firebaseplays
   2. cd backend/game-controller
   3. Copy the Firebase service account firebase-plays.json that you created in a previous step to backend/game-controller on the Windows game server.
   4. Edit config.js to include the API key, Project ID, Auth Domain, and Database URL values from your Firebase project.
   5. npm install
   6. node game-controller.js
3. Click the running game to ensure it has focus so game-controller.js can relay keystrokes into the game.

Sample BigQuery queries:

Most common game commands

SELECT command, count(command) FROM [social_gaming.game_commands] GROUP BY command

Commands by city

SELECT city, count(*) AS count FROM [social_gaming.game_commands] GROUP BY city

Commands by time of day

SELECT
  STRFTIME_UTC_USEC([timestamp], '%I %p') AS hour, count(*) AS count
FROM
  [social_gaming.game_commands]
GROUP BY
  hour