I am new to Speckle, and I am trying to use the Power BI visual to connect BOQs, schedules and cost estimates done in estimation tool with the 3D model.
My workflow relies only on IFC models export /import (export from authoring tools to estimation tool and then export from estimation tool to speckle)
I am using the latest release for the visual and power BI connector.
I have few questions in regards to the limitation of the size of the models.
Do paying users also have the 100 MB size limitation when uploading IFC files or is it only for free users? Is there a way to batch upload IFC models?
What is the limit of the model size that the viewer can handle?
I have tested it with a small model (the query table had about 16K rows) and it worked fine. but for a larger model (had to append multiple queries together and the number of rows is about 80K) it shows only part of the model and it is not responsive or shows the wrong elements when I use filters to limit the section of the building that should be visible.
I’d appreciate any tips on how to handle large models
Below are snapshot of the model and the visible parts in the power BI
For the limits/pricing question, it is new, and @benjavo could give you a better response for it. I will take over the Power BI related questions.
We have introduced a new way to get the data from data connector to visual with the next gen powerbi and it currently doesn’t support multiple models. But this is the project that we are currently working on to enable model federation. Our plan to have it in 2 weeks and I will let you know here when it is done.
Power BI visual can handle same size of data as viewer because they are same behind the scenes. The only limitation is the data storage limit of Power BI in visuals once we want to save this data into visual to be able to publish report into app.powerbi.com. But as a rough answer we managed to handle ~1GB Speckle data in our performance tests before.
In addition to @oguzhan’s answer, the 100 MB limit applies specifically to IFC file uploads, not to the data you can publish to Speckle. You can utilize our connectors to publish data directly from host applications instead of exporting IFC files, uploading them, and then loading them into Power BI. This approach allows you to bypass the 100 MB upload limit and provides enhanced data extraction capabilities without relying on IFC exports.
Let me clarify the workflow a little bit, structural models come from Tekla structures (IFC model), In the estimation tool (has no connector for speckle) I add to the model classification codes, costs and schedule data and export my power BI reports, I then export IFC file with only the classification data to keep it as light as possible. (I use the classification codes to connect both reports and the 3D model on the power BI side).
In this test project the IFC file was about 800 MB so I had to split it to 12 smaller IFC files to keep them under 100MB, after uploading I pulled all the 12 queries to power BI and appended them resulting in an 80K row query.
My understanding is that most power BI visuals can only handle ~30K row. Is it the case also for the Speckle visual? this is the error I get on the power BI side
I am thinking of two work arounds but not entirely sure if any of them will work
Push the classification data back to Tekla structures and upload the model to speckle through the connector- this solves the model size problem but not the visual performance problem. (also I was reading about issues accessing data on both assembly and part level for models coming from Tekla Structures).
If the issue is related to the 30K row in the query I am thinking to create 3~4 queries all under 30K row and some kind of measure/parameter so that the visual data can be controlled dynamically through a slicer, this way the visual never attempts to load the whole model at once.
The 30K limit should not be an issue for Speckle’s 3D Visual, as I have tested it with datasets containing more than 30K rows, and it performed without any problems.
I recommend publishing your data directly using our Tekla connector. You can also import estimation tables in Excel, CSV, or other formats supported by Power BI. Once you identify a field to link these two data sources, you will have a functional prototype. This approach will eliminate the need to import estimation data into the Tekla model.
I created a video tutorial on extending Revit models with Excel data using our legacy Power BI connector. Principles shown apply to the next-generation connector as well.
If you have trouble putting together a POC, please share sample files so we can better help you.
I understand that the model is large but I have no idea what is its size on the Speckles server. Also there is no way to filter only the needed properties to be transferred to Speckle from Tekla to trim down the size a little bit.
To conclude using IFC files the model loads partially, using the connector the power BI crashes.
I’d be happy to share privately the link to the model and the Power BI reports if you think that can help solve the issue.
