Python visual is not accurate

[Anonymous]

The matplotlib plot that shows up in Power BI, is different from that one, that the same code generates in a jupyter notebook.

Please help.

[Anonymous · Accepted solution]

Thank you @Pragati11  and @amitchandak .

The issue was with the sequence of columns. Power BI sends the columns in the alphabetical order, and not in the order in which the columns get uploaded.

Thanks again for your prompt support.

[Pragati11]

Hi @Anonymous ,

Can you share some screenshots here please with the following:

1. Screenshot of scripts used in Jupyter and Power BI
2. Screenshots of graph that is generated in Jupyter and Power BI

Thanks,

Pragati

[amitchandak]

@Anonymous , can share some sample images?

[Anonymous]

---

@amitchandak wrote:

@Anonymous, can share some sample images?

---

Thank you for the prompt response Amit.

This is the code:

import pandas as pd

import pickle

import matplotlib.pyplot as plt

enq_loaded_model = pickle.load(open('C:ENQUIRY.sav', 'rb'))

y = enq_loaded_model.predict(dataset)

plt.plot(y)

plt.show()

First one is from Power BI, and the second one from jupyter notebook

[Pragati11]

Hi @Anonymous ,

I can see the difference clearly in the y-axis scaling.

In Jupyter your y-axis range is till 2200, thereore the graph shows a different pattern.

In Power BI your y-axis range is till 2950, therefore graph is different.

In your pythin script when you are generating plot in Power BI, you will have to define range for your y-axis. Try giving this range similar to that in Jupyter notebook. I am sure you will get the same graph.

Thanks,

Pragati

[Anonymous]

Hi @Pragati11 ,

Thank you.

It didn't work. First one is from Power BI, and the seocnd one from jupyter notebook.

It didn't work.

Code:

import pandas as pd

import pickle

import matplotlib.pyplot as plt

enq_loaded_model = pickle.load(open('C:/ENQUIRY.sav', 'rb'))

# predict the enquiry

y = enq_loaded_model.predict(dataset)

plt.ylim([500,2500])

plt.plot(y)

plt.show()

[Pragati11]

Hi @Anonymous ,

Can you make sure that under this visual in Power BI, the fields that you have used - they are set to Don't Summarise?

This could be the reason why you don't see anything in Power BI for the same python script.

Thanks,

Pragati

[Anonymous]

Checked again, @Pragati11 . Each one of them is set to "Don't Summarize".

[Pragati11]

Hi @Anonymous ,

Is it possible to share your pbix file?

Thanks,

Pragati

[Anonymous]

Hi @Pragati11 ,

Hope you can access the files from here.

Dropbox 

[Pragati11]

Hi @Anonymous ,

Yes I can access them. Let me check and get back to you on this. 🙂

Thanks,

Pragati

[Pragati11]

Hi @Anonymous ,

I get following error on your script:

Is it working at your end?

[Anonymous]

Hi @Pragati11 ,

Must be something with the settings. It is working just fine. Please refer to the attached screenshot.

[Pragati11]

Hi @Anonymous ,

I am not sure. there doesn't seem to be a setting issue.

I tried a case at my end by using xlim and it seems to work just fine at my end.

See below my original chart:

Now I set a limit on my x-axis and I get the following:

The only different thing I have used in my code is ax = plt.gca() to get the current axis and then apply limit on the axis range.

Can you try adding that in your python script and check it?

Thanks,

Pragati