Building the BI system and analytics capabilities at the company based on Reality Games example
1 like297 views
The document presents a guide to integrating BigQuery with Firebase for app analytics, detailing the setup process and benefits of having accessible data. It describes the creation of a daily KPI table to streamline data analysis and improve reporting efficiency, addressing common challenges faced by analysts. The session includes step-by-step instructions for implementation and highlights the value of customized reporting for enhanced business insights.
Building the BI system and analytics capabilities at the company based on Reality Games example
- 1. Adam Longhorn Head of Analytics Reality Games adam.longhorn@reality.co https://pl.linkedin.com/in/adamlonghorn Making life easier and delivering insight faster: Simplifying App analytics and reporting in a few easy steps with Firebase and BigQuery
- 2. Award-winning f2p games using real-world data from across the globe Launched 2015 Launched 2016 Launched 2017 Launching 2020 Coming 2021 Coming 2021
- 3. ⓘ Start presenting to display the poll results on this slide. How long does it take to perform a 'typical' data analysis request from your team member?
- 4. ⓘ Start presenting to display the poll results on this slide. When would your team member 'typically' like to have their analysis request delivered?
- 5. What would be helpful for analysts: ➢ Accessible data, at the appropriate granularity ➢ Simple data structures so analysts (and quasi-analysts) are able to use it easily ➢ Relevant business metrics and dimensions that make it truly useful instead of just ‘nice to have’ data A challenge many analysts face (including myself): ➢ Business requests for analysis come in fast and business decisions get made even faster ➢ Our businesses generate a ton of data. A lot of that data not readily available to support the pace of the business
- 6. Solution: The ‘Daily KPI’ table ➢ An easy-to-use BigQuery table ➢ Granulated at the user/day level ➢ Covers 80% of ad-hoc analysis requests ➢ (Bonus) Connectable to Google Data Studio reporting SQL Tools:
- 7. Agenda for today: ➢ Integrate BigQuery with Firebase (assumes we all have Firebase SDK installed) ➢ Set-up SQL data-pull in BigQuery and schedule data-extraction job (step-by-step instructions and tools are given -- making your life easier!) ➢ Review what we have done, takeaways, and final thoughts Estimated time to implement yourselves: ● 1-min to integrate BigQuery ● 1-day of rest ● 30-min to build and schedule the jobs
- 8. ⓘ Start presenting to display the poll results on this slide. How familiar are you with Firebase and BigQuery?
- 9. Navigate to Firebase console, open your Firebase project and navigate to the BigQuery ‘Integration’ manager Select your project1 Project settings2 BigQuery3 Next window Next window
- 10. Toggle ‘Google Analytics’ integration Side note -- Other BigQuery integrations available: ● Crashlytics ● Predictions ● Performance Monitoring ● Cloud Messaging ...So much to analyze! Toggle this!1
- 11. Great job! Firebase is now sending raw analytics event data to BigQuery You have just opened up a new world of possibilities w/ BigQuery! ….. …. But what now?
- 12. Take a (1 day) break. You deserve it! AND You need to give Firebase time to export the raw data
- 13. Navigate to GCP platform, select your project and initialize BigQuery Scroll-down for BigQuery Select your project1 2
- 14. Navigate to your project and create dataset ‘reporting’ Highlight your project Create dataset Name: ‘reporting’ Create dataset Pop-up window 1 2 3 4
- 15. Paste in 1st SQL query (link) into any text editor Will look similar to this Point the 1st SQL query to your project’s analytics events dataset Find this string within the query: <analytics events dataset> ...Replace with: your project’s events dataset (ex: ‘analytics_201716494’) ….For each instance of <analytics events dataset> in the query 1 2 3 4 Link to 1st SQL query: https://bit.ly/DailyKPIs1
- 16. Paste modified 1st SQL query into editor, run and save as BQ table named ‘daily_KPIs’ Paste query here Run query Save results Clear editor Name table ‘daily_KPIs’ in the ‘reporting’ dataset Pop-up window 1 2 3 4 5 6 Save
- 17. Paste in 2nd SQL query (link) into any text editor Point 2nd SQL query to your project’s analytics events dataset Find this string: <analytics events dataset> ...Replace with: your project’s events dataset (ex: ‘analytics_201716494’) ….For each instance of <analytics events dataset> in the query (Exactly like we did with the 1st SQL query) 1 2 3 4 Link to 2nd SQL query: https://bit.ly/DailyKPIs2
- 18. Paste modified 2nd SQL query into editor and schedule query to run automatically every day Paste query here Create new scheduled query Clear editor Pop-up window Schedule it! 1 2 3 4 5 Fill in required data
- 19. Nice job! You are finished!
- 20. But wait! ….. What did I just create?
- 21. Scheduled job in BigQuery Extracts raw analytics event data from previous day IAP First opens sessions transforms raw data into daily summary for every user that logs into your app User info Appends user data to ‘daily_KPIs’ table on BQ to keep running history reporting.daily_KPIs Your Project space on BQDaily user summary
- 22. FIELD TYPE EXAMPLE DESCRIPTION event_date STRING 20200523 YYYYMMDD format for the day’s summary user_pseudo_id STRING 9544015614101dfce88afd6cf2ac6eb5 Unique identifier (to your project) for each of your app’s users country STRING United States Long form of user’s country platform STRING ANDROID ANDROID or iOS sessions INTEGER 3 Number of total sessions generated by user in given day seconds FLOAT 350 Total number of seconds user spent within your app on given day first_open INTEGER 1 1 = this is the device’s first time using your app. 0 = not first time payments INTEGER 2 Number of IAP user made on given day revenue FLOAT 9.98 Total amount of gross IAP revenue from user on given day (in USD) ‘daily_KPIs’ table: details
- 23. How your life has become easier: super quick analysis What was MAU, by country, in March? What percentage of new users converted to payers on D0 in the first week of February? What is the average session length by platform? What was our ARPU in January?
- 24. How your life has become easier: Customized/Flexible reporting with Google’s Data Studio Built-in connectors reporting.daily_KPIs Your Project space on BQ Your Logo Here!
- 25. How your life has become easier: Framework for more building a more sophisticated ‘daily_KPIs’ table You can modify the queries provided earlier by ... ● adding more dimensions (ex: mobile device or brand, app version, etc.) to group and filter by ● metrics from customized analytics events (level-ups, tutorial completes, etc.) that you would like to report on Enhance your own ‘daily_KPI’ table to… ● best-fit your businesses needs ● make your job easier ● deliver insight faster
- 26. Thank You for your time today and Good Luck! Adam Longhorn Head of Analytics Reality Games adam.longhorn@reality.co https://pl.linkedin.com/in/adamlonghorn
- 27. ⓘ Start presenting to display the audience questions on this slide. Audience Q&A Session