NBA Rebounding Analysis with Alteryx - Web Scraping & Predictive Analytics

Scraping the Glass: NBA Rebounding Analysis Web Scraping and Predictive Analytics with Alteryx Greg Murray Hector Amaya

1. Use Alteryx to scrape BasketballReference.com A. Return a list of all basketball players in catalog B. Return season statistics for all players in 2000 or greater Objectives 2. Use the resulting data set to build a predictive model

Batch Macros Download Tool Key Concepts XML parse Tool Regex Tool and Regex Parsing Alteryx Predictive Suite

Step 1: Examine the webpage(s) that will be scraped Objective 1a: Collect list of Players from BR Catalog

Step 2: Examine the structure of HTML for the desired section Use your web browsers inspect function (right click Inspect) Objective 1a: Collect list of Players from BR Catalog

Basic structure for HTML tables Column 1 Column 2 Column 3 Data 1 Data 2 Data 3 <Table> </Table> - Table Tag - Table Heading Tag - Table Row Tag - Column Headers Objective 1a: <thead> <tr> </tr> </thead> <tbody> <th>Column 1</th> <th>Column 2</th> <th>Column 3</th> Collect list of Players from BR Catalog - Table Row End Tag - Table Heading End Tag - Table Body Tag </tbody> <tr> </tr> <td> data 1</td> <td> data 2</td> <td> data 3</td> - Table Detail (cell) -Table Body End Tag - Table End Tag

Step 3: Use Alteryx to retrieve the HTML Pass first URL to download tool using a text input tool Use the Text columns tool parsing the HTML into rows Objective 1a: Add record IDs to rows Collect list of Players from BR Catalog

Step 4: Isolate code containing table data Challenge: We need to identify the table dynamically Table can be located on different lines on each page Table can have different # of rows depending on # of players Objective 1a: Collect list of Players from BR Catalog

Step 5: Isolate and parse the table headers Isolate table headers by reusing the section of tools we used to isolate the table Objective 1a: Parse the table headers using an XML Parse tool Collect list of Players from BR Catalog

Step 6: Isolate and parse the table detail Isolate table detail by reusing the section of tools we used to isolate the table and table headers Objective 1a: Collect list of Players from BR Catalog

Step 6: Isolate and Parse Table Detail Parse the table detail using a combination of XML Parsing and Regex Objective 1a: Collect list of Players from BR Catalog

Step 7: Union headers to detail Use the Autoc0nfig by Position option Ensure the header data stream in first in the Output Order Objective 1a: Collect list of Players from BR Catalog

Step 8: Validate the output Compare data in the workflow to the webpage Objective 1a: Collect list of Players from BR Catalog

Step 9: Convert the workflow into a batch macro Add a Control Parameter tool to Text input tool at the beginning of the workflow Objective 1a: Collect list of Players from BR Catalog Add a Macro Output tool the end of the workflow

Step 9: Convert the workflow into a batch macro Save the workflow as a Batch Macro into your macro repository If necessary add a new macro repository through the user settings Objective 1a: Collect list of Players from BR Catalog

Step 10: Create a new workflow using the new macro Pass the macro the player catalog URLs to download and parse all of the Player Catalog webpages Objective 1a: Collect list of Players from BR Catalog

Step 1: Examine the webpage(s) that will be scraped Objective 1b: Return Season Statistics for individual Players

Step 2: Examine the structure of HTML for the desired section Use your web browsers inspect function (right click Inspect) Objective 1b: Return Season Statistics for Individual Players

Step 3: Use Alteryx to retrieve the HTML Pass first URL to download tool using a text input tool Use the Text columns tool parsing the HTML into rows Objective 1b: Add record IDs to rows Return Season Statistics for Individual Players

Step 4: Isolate code containing table data Challenge: We need to identify the table dynamically Table can be located on different lines on each page Table can have different # of rows depending on seasons played There are multiple tables on a player s page Objective 1b: Return Season Statistics for Individual Players

Step 5: Isolate and parse the table headers Isolate table headers by reusing the section of tools we used to isolate the table Objective 1b: Parse the table headers using an XML Parse tool Return Season Statistics for Individual Players

Step 6: Isolate and the parse table detail Isolate table detail by reusing the section of tools we used to isolate the table and table headers Objective 1b: Parse the table detail using a combination of XML Parsing and Regex Return Season Statistics for Individual Players

Step 7: Union headers to detail/Append URL suffix Use the Autoc0nfig by Position option Ensure the header data stream in first in the Output Order Objective 1b: Return Season Statistics for Individual Players

Step 8: Convert the workflow into a batch macro Add a Control Parameter tool to Text input tool at the beginning of the workflow Objective 1b: Return Season Statistics for Individual Players Add a Macro Output tool the end of the workflow

Step 8: Convert the workflow into a batch macro Open the Interface Designer (CTRL +ALT + D) Change Batch Macro Output mode to : Auto Configure by Name Objective 1b: Return Season Statistics for Individual Players Save the workflow as a Batch Macro into your macro repository

Step 9: Use the batch macro to parse the player pages Open a New workflow Use the Player Catalog list as your input data Concatenate https://www.basketball-reference.com with the Player URL suffix Pass the Concatenated URL into the batch macro Limit the number of records per run using sample or filter tool Objective 1b: Return Season Statistics for Individual Players

Regex : www.regexr.com Batch Macros: https://community.alteryx.com/t5/Live- Training/Live-Training-Build-Your-First-Batch-Macro/m-p/52900 Alteryx Tool Mastery: Download Regex Resources