This document discusses building self-service analytics through data APIs. It defines self-service analytics and embedded analytics. It outlines three ways to provide self-service analytics: using third-party BI tools, building your own stack, or doing nothing. Building your own stack is challenging due to dependencies, APIs, and data models. Large companies provide extensive data platforms through predefined reports, dashboards, data APIs, and report designers that serve large datasets through a single data model. A "silver bullet" is a single, schema-less data API that can query any data, providing multiple use cases through a common access point and data model with transparency.
4. Analytics for Customers - Why?
• Better Productivity based on
Analytics
• Lower Operational Cost
• Improve Customer Retention
• Gain Competitive Edge
• Generate Additional Revenue
“Analytic Adopters
are x3 more likely
to see and increase
in revenue of
greater than 20%”
5. Self-service Analytics: Define
• Dashboards & KPIs
• Predefined Reports
• Scheduling
• Designing Reports/Report Writer
• Power-Pivot
• Filtering
• Data API
• Data Exports
• Design/Create
Own Data Context
Customer 1
Customer 2
Customer N
…
6. What is Embedded Analytics?
Embedded analytics is the integration of business intelligence
and reporting capabilities directly into an application.
“Analytics” mean:
• Dashboards & KPIs
• Designing & Running
Reports
• Self-Service
“Embedded” mean:
• Infusing 3rd party BI tool
into your Product/Platform
• White labeling
7. 3 ways to get to it:
• Using 3rd party BI Tools and
Solutions (Embedding)
• Building your own stack
• Forgetting about all of this
Decision Drivers:
• Customer needs
• Price
• Legacy
• Backlog
11. Scaling
• Licensing
• Deployment
• Pipelines
• Intermediate storages
• Automation
• More Data sources
• BigData?
• Management and ownership
12. Conclusions
• Good for Small/Mid-size projects
• Not so trivial for Bigger scale
• Could be expensive
• Licensing
• Development is still needed
• Difficult to Operate
• Not all use cases covered
15. We can extend it
Pre-Defined Reports
(s02-e01)
- More open source (Plotly, Highcharts..)
- Microservices!
- Interesting…
(but nobody cares)
Dashboards
16. Let’s try to finish…
Pre-Defined
Reports
(s03-e04..)
- JS gurus will help…
- Maybe GraphQL?
- Not funny anymore…
Dashboards
Report
Designer
Blue dream:
External
{REST API}
17. End results: “As usual”
• Exciting & Interesting Journey
• Long way & Expensive
• Difficult to Maintain
• Dependencies!
• Swamp in APIs and Data
• Many data models
• Not all use cases covered
• Report designer?
• Data API
19. How about big companies? How they
are solving?
• Provide rich extensive data platform:
• Predefined Reports, Dashboards, Data APIs, Report Designer, Data
Exports…
• Thousands of data slices?
• Keep one data model and standard naming?
• Serve large datasets?
• Manage dependencies?
• Have full transparency?
26. More Use Cases
• Self-service analytics module within your solution
• Data API for your external partners and internal
development teams
• Smooth migration to other data storage technologies
• Update data structures based on usage metrics
27. Data API Implementations and
Examples
• developers.google.com/analytics/devguides/reporting/core/v3/refere
nce
• https://ga-dev-tools.appspot.com/request-composer/
• tech.yandex.com/metrika/doc/api2/api_v1/examples-docpage/
• http://cube.dev/ (Slack)
• https://www.peekdata.io/data-api.html
• http://reportbuilder.peekdata.io/
• https://gitlab.com/peekdata
29. Takeaways
• It depends
• Many BI tools and solutions around, but…
• Toolset, for building your own Data Platform is complete
now
• Nothing comes for free
Enjoy development - building, embedding or whatever!
30. Make your data easily
accessible, every time and
everywhere
Thank You!
www.peekdata.io