#91 End-to-end analysis of participants race Cross de Invierno 2023 with Microsoft Fabric (and Power BI)
In this old post I wrote about the run Cross de Invierno 2023 in Madrid I which I participated, which included an analysis of the finish-times, in Excel.
In this new post I used the data of the this run again, but this time for an analysis of the participants of schools and athletic-clubs, and now I used Microsoft Fabric.
Fabric is Microsoft's quite new SAAS data analytics platform (it went GA in Nov.2023). It's a suite of tools to facilitate end-to-end analysis of data, and includes Power BI, One Lake, Data Factory (Data Pipeline, Dataflow), Synapse Warehouse and Lakehouse, Spark Notebooks etc. Fabric is made for 'personas' as Data Engineer, Data Scientist, Business Analyst etc., for which there are different built-in workloads ('experiences').
For more info about Fabric, see e.g:
https://learn.microsoft.com/en-in/fabric/get-started/microsoft-fabric-overview
https://youtu.be/J4i5lcROJcs?si=ZPo-mgnmQx8eRPg2
NB: in min. 19:16 there's a good picture of the Fabric architecture.
Evolutionary History of Microsoft Fabric - Spreadsheets to Lakehouse
The Power BI report that I made has these race-participants stats:
The results of the race are published in several PDF-files, and for my analysis, I used 2 files, with the results of all men and women, so all adults (categories senior and junior). I converted these PDF-files to Excel and uploaded them in my Onedrive. I then used a Data Pipeline with 'activity' Dataflow Gen2 (with PowerQuery) to load the data from the Excel-files into a (Synapse) Lakehouse:
A Lakehouse is a data asset of OneLake, the central storage place of all data loaded in Fabric.
All data in OneLake is stored in the Delta Parquet format, an open data format that can be read by all engines in Fabric (e.g. Power BI, using Direct Lake mode), which looks like this :
NB:
-note that the files are in the OneLake-folder, the 'OneDrive for data', see:
https://youtu.be/4NbuG1JBq60?feature=shared
-for more info about Onelake, see:
https://radacad.com/what-is-onelake-in-microsoft-fabric-and-why-you-should-care
-for more info about Delta Parquet format, see:
https://data-mozart.com/parquet-file-format-everything-you-need-to-know/
I made a Notebook to query the tables that I created in Onelake (Lakehouse), using the pyspark Python-library:
And another notebook with the sempy-Python library. with which a PowerBI semantic model can be queried, using DAX: 
I created in a Fabric Warehouse the view RunnersAll, to combine the data from the 2 tables MenAll and WomenAll, with this SQL (run in the LakeHouse's SQL-endpoint):
In Fabric there's a Visual Query Editor (PowerQuery) that can generate the SQL for you, here an example to get a list of all teams of both men and women runners:
And the generated SQL:
Note that in this (generated) SQL "DISTINCT .. UNION ALL" is used, where I would have used UNION (which filters out the duplicates, no separate DISTINCT needed).
I made some other (dimension-)tables that I then added in this Power BI Semantic Model (star-schema):
And this model was the source for this Power BI report:
The table in this dashboard (PBIX-report) with the details of the runners is a long list to scroll through.
A better way to show (long) lists are paginated reports (RDL-report). There are 2 ways to make a paginated report:
*1: in the Power BI Service, using the semantic model, and a query in DAX:
*2: in Power BI Reportbuilder, using the SQL-endpoint of Fabric Lakehouse/Warehouse, and a query in SQL (report shows top 3 runners per category):
The query for the dataset in this report can be made in SQL or the Query Designer (that generates SQL):
NB: for a good post about PBIX-reports ('dashboards') vs RDL-reports ('classic' paginated reports), see:
https://sqlserverbi.blog/2022/02/04/when-to-use-paginated-reports-vs-interactive-power-bi-reports/
To see if the schools that participated in the run were all coming from my neighbourhood (Ciudad de los Poetas) where the race was held (park Dehesa de la Villa), I made this PBIX-report:
MS Fabric supports version-control in Git, see column Git-status 'Synced' :
I worked on this project in Dec.2023 and Jan.2024 (with a free Fabric trial license), so shortly after
MS Fabric went 'GA', although actually it was still a beta-version back then... But what is the status now, in Jun.2025, is MS Fabric 'production-ready' now? It looks it still isn't:
https://www.brentozar.com/archive/2025/05/fabric-is-just-plain-unreliable-and-microsofts-hiding-it/
Microsoft Hates Greg - MS Fabric Performance & login issues
Microsoft Fabric: 10 Reasons It’s Still Not the Right Choice in 2025
And is Microsoft's idea to bundle all their data-tools (as Power BI and Azure Synapse) in one suite (Fabric), a good idea? Some more critical opinions:
https://www.linkedin.com/pulse/microsoft-fabric-repeating-sins-past-greg-deckler-mrrae
Curbal - I can explain Microsoft Fabric with ONE WORD
I googled with "MS Fabric (customer) success stories 2025" and "MS Fabric market adoption 2025", and found e.g.:
https://www.reddit.com/r/MicrosoftFabric/comments/1jz4dxz/microsoft_fabric_success_stories/
https://www.microsoft.com/en/customers/story/23991-lumen-microsoft-fabric
MSHGMQ - Microsoft Fabric and the Uncanny Valley
As they say in this post:
"Only time will tell if Fabric stands the test of time against competitive technologies.[e.g. Databricks or Snowflake]."
If you enjoyed this post and want to make a small donation, you can do this with the Paypal Donate button (at the top right of this post) or via BuyMeAcoffee .
Downloads
PowerBI report Schools of runners
Excel-export Paginated report Runners
Notebook sempy