

Power BI is absolutely amazing. So simple to use, so beautiful, so quick and powerful. I’ve only used it for about 6 weeks, and have very quickly built a BI system for an organisation struggling with data, and the responses where between “great” and “wow”.
For me, having built BI systems using OBIEE, Cognos, Tableau, and even the old Oracle Discoverer back in the late 90s, the coolest thing about it was the fact I could deliver in hours.The business model is amazing – everyone has an account even before hearing about it, and pricing is very attractive.
There are allot of blog posts and videos about the greatness of Power BI, but I actually want to discuss my wish list for making it even better. Some of these items are already in the works and will be available in the next few months. Some of the items may have workarounds which I’m not aware of. Hey – I’ve only got 6 weeks of experience, right ?OK, so here’s my wish list, sorted by importance :
1. Incremental Refresh
Once a dataset is published, we can schedule refreshes, up to 8 times a day. The problem is – a refresh may take a while, depending on the dataset size, the data types, latency, and the source databases(this is crucial by the way – I’ve seen loads from SQL Server run 10 times better then from Oracle). A dataset the size of a few gigs, from a slow source, can take 30,40,50 minutes to refresh, even more. What if we could just load the new or changed records? A definite number one on my list.
2. Run on-premise
Some organisations simply cannot load data to the cloud. I’m now starting to build a BI solution for a company who legally cannot load user information to a cloud, so we are very limited in what can be done with Power BI. There are workarounds like direct query(which does not allow some of the features, and latency is a big concern), or publishing to a Pyramid Analytics server on premise. Haven’t tried that yet. Microsoft is already working to enable partial on prem support as part of SQL 2016. I’m not sure yet what’s in and what’s out though.
3. Save filters
Plain and simple – I want to save different sets of filters so I can reuse them the next day. And the day after.
4. Use the same dataset for several reports
The most important part of creating a power BI report is modelling the dataset: choosing the sources, defining relationships, renaming fields, changing data types, creating calculated columns and measures, etc. Once done, beautiful reports are created easily using Power BI desktop. But, a dataset and a report get published together. If I want to create two different reports based on the same dataset, I can only do it in the service, which is limited comparing to the desktop tool. Since I have different users in different groups who may need to see different parts of the data, I found myself copying a dataset 3 times. It may be fine to maintain 3 copies, but I wouldn’t dream of doing that with 10. I’d like to see Microsoft separate datasets from reports. I’m sure I’m not alone.
5. Customize Field names
A field name is determined on the dataset level. Then it can be used in many visuals inside the report, but it’s always called the same name, Sometimes I want to be able to call the same thing two different names in two different visuals. Cannot be done at the moment.
6. 10Gb size limit
Every pro user can keep up to 10Gb of data, including content packs which are shared with them. That’s not bad, especially given the great compression of the datasets. However, I’d like to be able to pay more and get more.
7. Refresh interval
Haven’t run into a problem yet but I’m sure I will. 8 times a day per dataset may not be enough in some cases. If we want to use Power BI for Operational Intelligence, we need to be able to refresh every few minutes.
That’s it, really. For now anyway. But even if none of that gets done, it remains my favourite BI system.
How do you feel about my list ? what’s yours ? Have you worked around some of these issues? Please comment.
Cheers.