1. Action, get started todays
2. Quick Wins, aim for quick small wins and ITERATE
The source data may be manually compiled by the author, usually for testing purposes.
|Connected to the source system
The source data is queried from the proper source system.
|Efficiency and optimization
Queries are optimized for efficiency, such as query folding when querying the source system.
The dataset is manually refreshed in Power BI Desktop by the dataset author.
The dataset is scheduled in the Power BI Service, so it refreshes automatically.
The dataset refresh operation is optimized for incremental data refresh.
Remarkably similar to PivotTable Data in Excel, where you used lookups in the source (Most likely Excel or current source).
Fact-Dimension, Star Schema, Calendar Table.
|Optimized Storage & Relationship
Wide and Short Dimension Table & Skinny and long Fact Table Read More
Get started with drag and drop, but highly recommend writing Explicit Measures. Read More Sales = SUM( ‘Invoice Data'[Revenue])
|Calculate/ Time Intelligence
Function: Calculate/Filter, Time Intelligence (DateAdd, DatesYTD)
Introduction to X Functions,
Filter Context etc.
DAX/Measure Optimization, unlocking the power of Calculate/Filter/Iteration/Table Functions, Composite Models
1) DAX Studio 3) Bravo
2) Tabular Editor
|Follow K.I.S.S method
Simple Tables/Matrix and Charts. Let the interactivity of Power BI along with Drill Down do the magic, just getting what you have replicated.
|Move beyond The Grid
Drill Through, Information canvas, attention direction flow, information positioning, color contrast, conditional formatting.
|Sprinkle on Top/Insight
Bookmarks, Navigation, advanced custom visuals (Inforiver, Zebra BI).
Every Custom Visual has its own learning curve.
Being able to synthesize information into simple nuggets. IBCS
1. Date Prep: Power Query/M language/Get Data to transform and Load
2. Modeling: Model/DAX/Relationship Semantic Layer.
3. Visualization: The Art/Science of Story Telling.
Here goes nothing. I’ll try to lay down my method to the madness for success/adoption/efficacy with Power BI.
|Amaze yourself that you built an interactive dashboard in less than a couple of hours.
|1. Make sure you build/convert one of your current reports or dashboards
2. Do not share this version
|Amaze/tease your consumers and collect feedback.
|1. Ensure your report is presentable and learn to embrace the blank space, DO NOT PUT too many things just because you can
2. Ensure your report is presentable and learn to embrace the blank space, DO NOT PUT too many things just because you can: **Win people over with the power of familiarity, Use Power BI Secret weapon Analyze in Excel.
Music and Emotions in the Brain : Familiarity Matters
3. I know your sources are still manual, so not that much capacity has opened up, but you’ll get the boost from the win.
|Ease of communicating messages and unlock dynamic calculations for new capabilities.
|1. Move beyond VLOOKUP/Index-Match to the power of relationship and introduce the Calendar table.
2. Significant benefit by having Time Intelligence and dimensionality.
3. Share it with your consumers, do not forget their adoption speed would vary and continue to use the power of familiarity in Analyze in Excel.
4. Still Manual Source, but Start investigating access to source/auto-refresh capabilities.
|You should completely replace and replicate your old report.
|1. Ensure you have parity and superiority to the old report.
2. Cut off the old report to open capacity.
3. Models and Visualization will continue to be improved; however, you don’t want to get stuck in doing both methods.
4. At this point your Effective way should first go for data modeling and then visualizations.
|As we are Effective now, focus on Efficiency
|1. Date Auto Refresh Strategy: Best Solution: Direct access to source ERP/CRM etc. or any available Data Warehouse.
Means working with IT to get access, it’s hard but you must communicate with IT frequently to get access.
If that fails, focus on if there are automated reports that is being sent out, it’s a form of batch jobs.
You can store them preferably in SharePoint Folder and have Power BI refresh from there.
You can use Flow/Power Automate to connect and get output automatically as you want.
2. Why did we not evaluate Automation earlier?
Primarily because it takes a lot of time and is a bureaucratic process.
Honestly, your consumers DO NOT CARE as you have nothing to show them.
It is particularly important to achieve automation for sustainability, but this is not a battle you pick first if you want to win the war.
Also, if you do not free yourself up at this point, you cannot focus on bigger and more complex requests from your consumer.
|Utilize more advanced design/modeling capability to solve questions that you couldn’t before.
|1. Higher ROI than visualization as answers more complex questions.
2. This step might come early if you have hundreds of Millions/Billions of records.
3. Do not complicate it if there is no value.
|Focus on visual finesse and wow your consumers.
|1. At this iteration, you have answered the questions but now you can focus better on storytelling to effectively communicate the insight.
|The journey does not stop as there is always a new question to answer or a change in leadership that wants a different color…….
|1. Just remember what success really looks like, so, be open to change and adapt.
Purpose: Amaze yourself that you built an interactive dashboard in less than a couple of hours.
Purpose: Amaze/tease your consumers and collect feedback.
Purpose: Ease of communicating messages and unlock dynamic calculations for new capabilities.
Purpose: Should completely replace and replicate your old report.
Purpose: As we are Effective now, focus on Efficiency.
Key Points: Date Auto Refresh Strategy:
Best Solution: Direct access to source ERP/CRM etc. or any available Data Warehouse
Purpose: Utilize more advanced design/modeling capability to solve questions that you couldn’t before.
Purpose: Focus on visual finesse and to wow your consumers.
Purpose: For Large and complex Data Engineering, optimize load time and stress on system.
Iteration 9 & on:
Purpose: Journey does not stop as there is always new question to answer or change in leadership that wants a distinct color …