ML Made Simple

Exercise Part 3: Exploring Data for Machine Learning

Get ready to dive into the world of data! In this exercise, you’ll use historical bicycle rental details to train a machine learning model. The goal? To predict the number of bike rentals on any given day based on seasonal and weather patterns.

---

Step 1: Understanding the Data

• The Data’s Story: We’re using data from Capital Bikeshare to understand how various factors like weather impact bicycle rentals.

---

Step 2: Creating a Dataset in Azure ML

• What’s a Dataset? In Azure Machine Learning, a dataset is like a container for your data. It’s where all your model’s food (data) is stored.

---

Step 3: Getting the Data

1. Find the Data:
   • Check out the data here.
   • Save this data as a local file named daily-bike-share.csv. It doesn’t matter where you save it, just remember the spot!
2. Upload to Azure ML Studio:
   • Head over to the Datasets page in the Azure Machine Learning studio.
   • Datasets are like the maps of your data world in Azure ML.
3. Create a New Dataset:
   • Basic Info:
     • Name: bike-rentals
     • Dataset type: Tabular (think of a neat data table)
     • Description: Bicycle rental data (simple, right?)
   • Datastore and File Selection:
     • Use the currently selected datastore.
     • Browse and select the daily-bike-share.csv file you downloaded.
   • Settings and Preview:
     • File format: Delimited (like a neatly organized drawer)
     • Delimiter: Comma (the separator that keeps data in its place)
     • Encoding: UTF-8 (the language of the data)
     • Column headers: Only the first file has headers
     • Skip rows: None (we want all the data!)
   • Schema:
     • Include all columns except ‘Path’.
     • Review and confirm the automatically detected types.
4. Dataset Creation and Exploration:
   • After creating the dataset, open it and visit the Explore page. This is where you’ll see a sample of the data and get a feel for it.

---

Citation:

• The data used here is courtesy of Capital Bikeshare and follows their license agreement.

---

And voilà! You’re now ready to start analyzing the data. Who knew bike rentals could be so interesting? Next stop: building a model that can predict the future… of bike rentals, at least!