added improved jupyter notebook and corresponding csv files

[marvinbuss]

Purpose

• Added an improved Jupyter Notebook with csv file support and plotting in matplotlib.
• Added a connection to Power BI to simulate data streaming to Power BI.
• Merges all functionality into one notebook.
• Other small changes and improvements (better transparancy).
• Added csv-files and improved the naming scheme of files.

Does this introduce a breaking change?

[x] Yes
[ ] No

Pull Request Type

What kind of change does this Pull Request introduce?

[ ] Bugfix
[ ] Feature
[x] Code style update (formatting, local variables)
[ ] Refactoring (no functional changes, no api changes)
[x] Documentation content changes
[ ] Other... Please describe:

How to Test

• Get the code

• Install Anaconda and run the notebook.

• Test the code

• Just run the Jupyter notebook and verify the functionality.

What to Check

Verify that the following are valid

• Verify the correct functionality of the notebook. I tested it all, but please validate correct functionality.

Other Information

[msftclas]

All CLA requirements met.

[yingqunpku]

@TonyXing and @conhua could you help review Marvin's PR?

[yingqunpku]

@marvinbuss
In your notebook, I have noticed send_to_power_bi function. I assume there'll be a proper PowerBI setup needed in order to make that function work (with POWER_BI_REST_API_URL). What is the setup process? Is it possible to have a powerbi desktop project or a readme to guide users have their pbi setup ready to consume the api call?

[marvinbuss]

@yingqunpku
You can find a description in my repository (https://github.com/marvinbuss/MS-AnomalyDetector). I can add that to one of the Readme files in your repo.
I can also remove the feature, if you want me to.

[yingqunpku]

@marvinbuss a few comments:

1. I'm not able to try out the PBI function as the code fails on my local env
   TypeError: ufunc 'isfinite' not supported for the input types, and the inputs could not be safely coerced to any supported types according to the casting rule ''safe''
   error.txt
   The error has nothing to do with the PBI function though.

2. In Powerbi, there're 3 options when creating a streaming dataset: api, azure stream, and pubnub. I guess it should go with api the default value but you may want to clarify that with your readme.

3. read_csv("data/sample_hourly.csv") "data" dir does not exist in this repo. You may want to change all those to ../example-data/ OR create a folder and move your data files to the folder.

[marvinbuss]

@yingqunpku

1. I was not able to reproduce the error, but other users report the same issue (https://stackoverflow.com/questions/29329725/pandas-and-matplotlib-fill-between-vs-datetime64). I updated the notebook with the proposed solution (data['Date'].dt.to_pydatetime()). This should hopefully fix the issue with some pandas versions.

2. I updated the Readme File in my repository and added some more details to the instructions: https://github.com/marvinbuss/MS-AnomalyDetector

3. Sadly, I comitted the wrong file on Friday. This was already fixed in the first commit.

4. We should delete the JSON files in AnomalyDetector/ipython-notebook/ and update the file names in AnomalyDetector/example-data/. Currently, it is not very well structured and can be improved. We could also keep the files that I uploaded in this branch.