I am looking into Power BI as a solution for about 400 users. One of my goals is to have a lot of dashboard publishers and PBI is great at that vs Tableau due to cost, strong data management capabilities, Excel integration.
Where I am struggling though is the idea of doing data management within PBI and as a consequence, not pushing for a data lake strategy where all data is loaded in 1 database.
Currently my team compiles data from various data sources, blends it, cleans it, transforms it, using Alteryx and in the future Python. After compilation the data is loaded in Exasol. Then both Tableau or PBI can connect to Exasol and people can build visualizations as they like.
We are not a warehouse, all data is managed by business users, but it is done in 1 place and it is done in a very automated fashion.
Question is, are we moving away from data lake strategies with PBI? If yes, how do you plan to manage this from becoming a spreadsheet nightmare all over again (since there is no control on data)? Also, if yes, will PBI be able to refresh joins and transformations on a scheduled basis?
Those are some of the advantages I am after with the data lake:
- Control over data and 1 source of truth
- Automated data refresh with 0 manual intervention (daily, weekly, etc.)
- Complex transformations (if/else, etc.)
- Data catalogue so people know where to find the data for their analysis
- Training on 1 data catalogue
- Do data work 1 time instead of n times (1 for each dasboard creator)
