Steps to become data driven
1.1. All the information in a unified system
- All new data should be in a unified system, such as WikiSuite.
- Relevant data from legacy systems should be progressively migrated.
- If some information needs to remain in an external system, determine an interoperability mechanism, for example, with https://doc.tiki.org/Tracker-Tabular or (as a last resort): https://doc.tiki.org/PhantomJS-and-CasperJS .
- Even if you are still file-centric, at least you can stop emailing files back and forth and wondering who has latest version.
- Set up an Analytics system to start collecting data.
- You'll now have a search engine that goes through everything, (Manticore Search in the case of WikiSuite).
1.2. Information in proper format
- Use the appropriate format for all new content.
- Wiki pages instead of word processor documents
- Tiki Trackers or Tiki Spreadsheet instead of a file-based spreadsheet
- Pivot Tables
- Gantt Charts
- Use templates and guidelines to maintain coherence between sections, while providing flexibility to adapt to specific needs.
- With tracking of changes (for all types of data)
1.3. Reports and dashboards
- Identify your Performance indicators and build relevant reports.
- Web analytics.
- Kibana for all data in Tiki, which is made available via Elasticsearch.
- Tiki's Calculation Syntax
1.4. Automation and alerts
- Identify key issues and risks, and add alerts.
- When problems occur, ask yourself if this could have been prevented or would have been less severe had we known about it. If so, how can we set up an alert to mitigate future problems?
- Related tools:
1.5. Machine Learning
- Detect anomalies from usual patterns (such as security breaches, errors, opportunities, etc.).
- Find trends and patterns from complex data sets and try to predict future events.
This is on the roadmap for Tiki and WikiSuite: