topic Re: using Python pandas pus matplot lib for advanced visuals in Custom Visuals Development Discussion

using Python pandas plus matplotlib for advanced visuals

flapponi — Thu, 28 Oct 2021 08:48:27 GMT

Hi,

I am new in this community. I am transferring some python codes into a power BI project in order to share with my colleagues. I put together a code to plot the cumulative sum of certain curves with fill colors in between. I transfer it ito power BI and the result is not what it should be expected (the code run fine on my jupyter notebook however). see below the

result

expected (from jupyter notebook

maybe I am pushing too much the limits, however is there anything I should consider?

Re: using Python pandas pus matplot lib for advanced visuals

dm-p — Thu, 28 Oct 2021 00:06:17 GMT

Hi @flapponi,

Without seeing your code or data I can only speculate, but it's highly likely that your dataset is being aggregated to distinct values by Power BI, which is it's default behaviour (vs. tools like Jupyter, which take a dataset as-is and process row-by row). You can confirm this by clicking the visual header and selecting Show as a table, which will likely confirm that you have less total rows in your visual than in your dataset.

The simplest solution is to add a column to your dataset that will ensure Power BI creates the desired level of granularity for your data frame in Python. This approach is written up in longer form (with illustration of the concept) here; while this post focuses on a specific visual, the approach is portable for any visual that needs the lowest level of grain possible. We've also discussed this consideration for another tool external to Power BI vs. internally here, which again uses this approach and may help clarify further.

If this approach does not solve, you may need to provide further information on your dataset for us to help further.

Regards,

Daniel

Re: using Python pandas pus matplot lib for advanced visuals

flapponi — Thu, 28 Oct 2021 08:50:46 GMT

Hi ,

txs for yuor prompt answer! the dataset seems to be correct. I have plotted using a bar chart with continuous axis and I get the desired result (see image below). however this visual has a very annoying limitation as it seems not to allow to sort the axis value according to your prreference (ascending or descending).

Interestingly adn surprisingly, this is only allowed using categorical y axis type!?!? but this visual is not really necessary as there are way toomany categories in this way and I would eneed to downsample my data (the y axis is technically a continuous data type)

I'll see the best way to share a subset of my data and code for a test if you like to.

Txs again and regards

Fabio