Introduction to supervised machine learning in Elastic webinar
Tuesday, February 25, 2020
Wednesday, May 27, 2020 - updated customer_churn.ipynb for version 7.7.0
Monday, November 16, 2020 - updated for version 7.10 and eland 7.10b
Monday, September 20, 2021 - updated customer_churn.ipynb for version 7.14
Monday, February 8, 2022 - updated customer_churn.ipynb for version 8.0
Data and Python Jupyter Notebooks
Customer churn demo
- calls.csv
- customers.csv
- customer_churn.ipynb
Energy price forecasting demo
- energy_price_forecasting.csv
- energy_price_power.csv
- energy_price_weather.csv
- energy_price_forecasting.ipynb
Language Identification Demo
This demo requires analysis-kuromoji, analysis-nori and analysis-smartcn to be installed.
Console API calls are in lang_ident_console_cmds in this gist.
See blog for more information.
Kibana saved objects
- saved_objects.ndjson
Installation
This is how these notebooks were developed and tested (conda 4.8.1).
Create a local environment (in this directory):
python3 -m venv elastic_webinar
source elastic_webinar/bin/activate
Install requirements
pip install --upgrade pip
pip install -r requirements.txt
Run notebook server
jupyter notebook
Elasticsearch connection configuration
To connect to an Elastic Cloud Elasticsearch cluster add connection details to cloud_id.csv and credentials.csv.
cloud_id.csv should contain the cloud_id value shown in the Elastic Cloud deployment page (without \n)
credentials.csv should contain the password for the elastic user to connect to Elasticsearch
e.g.
$ cat cloud_id.csv
ml_webinar:ZXVyb3BlZXdlc3QzLmdjcC5jbG91ZC5lcy5pbyRlZDBmMTQ2M2I3MTM0NjIwOGRlMzAzZWJjYmIzNGY2MSQ5NWE2YzNhYTBiMzk0YzI0OWEwYzI2Y2JlOTQ2YWY2Ng4=
$ cat credentials.csv
234dsp8ae42KWdYDZkn9Sg29
To connect to a local Elasticsearch cluster change CLOUD = True to CLOUD = False
Open, explore and run notebooks
The notebooks summarise the demos performed in the webinar.
Note running these notebooks may take ~1hr on a modest Elastic instance