Machine Learning in Power BI

[@judycalvo5127]

This was exactly what I needed to integrate my machine learning scripts into my Power BI reports. Awesome content!

[@arnaudmoulis454]

Thank you! This is exactly what i needed to see

[@henriquetestai3748]

Thank you so so much! Excellent presentation skills and pace (and content!!)

[@286pankaj]

Great. Excellent Work. One Query- How can we do What if scenario here? For example if we increase salary what will happen on prediction

[@absentdata]

Each feature has a coefficient so you can get the coefficient for salary and see to what degree 1 unit of salary changes the prediction,

[@286pankaj]

@@absentdata Thanks for your reply. I do not have much knowledge on this, can you explain how the salary coefficient be achieved.

[@amelieogcharline]

Great, short presentation. However, you cannot use the labelencoder on the department like that. You need to have one column pr department containing a binary variable. Please be aware of this, you can use one hot encoding in Python for this.

[@chrisau2159]

This is really important otherwise the model will assume employees with a higher department value are weighted more.

[@xXKiLeRMoDeXx]

@@chrisau2159 agreed. Should have individual columns with binary values for each. Message me if you need the code for binary columns.

[@padraichayes3836]

@@chrisau2159 with binary will the model still not just do the same and weight based on the quantity (occurrence of each department)? Just starting with Python here

[@addendanalytics-microsoftg5620]

Great video. Very informative and clear

[@juangut4531]

Awesome! One question please. What if you wanted to make a prediction for new rows that aren't in the original dataset? Do you refresh the data set in power bi with the new info??

[@absentdata]

You would simply add new rows to the dataset . the model is trained and will make the predictions on the new rows in the dataset

[@juangut4531]

@@absentdata thanks for answering!!

[@1982cannabis]

@@juangut4531 Hi Juan, Have you been able to do that?

[@rose8968]

Great explanation. Helpful!

[@absentdata]

Thank you!

[@leleuf3]

Man, you're awesome!!!!Thanks!!!

[@MrBlazeTheNight]

Is it possible to share this dashboard with friends without them having to install python? if they have PowerBI

[@absentdata]

Hi...The answer is yes you can have them access the dashboard online through Power BI services. They will only need to interact with it. The data and code environment exist on your PC at that point

[@MrBlazeTheNight]

@@absentdata But if i am building a time series model that needs to be updated every month (i.e. new data comes in every month) is this possible to still do it through powerbi and that the client can interact with the new data?

[@mikkelns8287]

This was great, thank you!

[@TheStoryTeller2898]

I am not able to run my script completely. I am getting No module sklearn. I am getting stuck at the import part from sklearn. Please guide how to resolve

[@absentdata]

Try this video it should help you out: kzbin.info/www/bejne/e6W0iIudYr2inKs

[@weriksonettone9790]

All times error DataSource.Error: ADO.NET: Python script error. Traceback (most recent call last): File "PythonScriptWrapper.PY", line 2, in import os, pandas, matplotlib ModuleNotFoundError: No module named 'matplotlib'

[@manmeetsingh3036]

thanks for the great explaination

[@MrSandeep8727]

Thanks for the excellent presentation. Can you please share the data set used for this exercise? Thanks

[@absentdata]

www.absentdata.com/power-bi/python-machine-learning-in-power-bi

[@closed252]

@@absentdata I think this link is broken.. could you, please, send a new link??

[@eyalakhdhar520]

Hi, please i want to connect power bi to python, every time i have multiple errors

[@absentdata]

Here is a way for you to easily avoid this. I will make a video on this hopefully soon. www.absentdata.com/power-bi/avoid-python-errors-in-power-bi/

[@momosky9168]

Thanks you for video :)

[@deepaka3820]

Thank you, Donald Trump.. this was very useful 😇

[@absentdata]

Glad you enjoyed it!

[@ntrax8214]

At 10:56 I can't see how you load the new table. help!!

[@absentdata]

Sorry for the late reply . The last part of the code should produce a dataframe that you can open up

[@AnilKumar-bd2yt]

You are testing on your whole data 😂 you should only test on y_test or use cross validation