Tableau Baseball: Player Performance Visualization

A Tableau data visualization project built for Udacity's Data Analyst Nanodegree, course "Data Visualization in Tableau." The workbook explores performance characteristics of a dataset of professional baseball players, looking at how batting average, home runs, height, weight, body mass index (BMI), and handedness relate to one another, and is published as an interactive Tableau Story with two iterations (an initial version and a final version revised after peer feedback).

Data

The dataset is baseball_data.csv, containing one row per player with the following columns: name, handedness (R/L/ambidextrous), height (inches), weight (pounds), avg (batting average), and HR (home runs). It is the baseball player dataset supplied as part of the Udacity Data Analyst Nanodegree's Tableau course material. Per the project write-up, the raw file included 1,157 rows, some of which were duplicate or inconsistent entries for the same player name; these were investigated and cleaned as described below before the CSV was used in Tableau.

Method / What was done

1. Loaded baseball_data.csv into Tableau and built an initial scatterplot of batting average versus home runs by player name.
2. Found that two names (Bobby Mitchell and Dave Roberts) each had two rows with different underlying data, since human names are not reliable primary keys. Relabeled the duplicated names with (1)/(2) suffixes so each row plots as a distinct player.
3. Cross-checked for further duplicates using a height versus weight scatterplot, which surfaced physically impossible entries (for example, a player recorded at over 12 feet tall). Removed one true duplicate row (Mike Brown) from the CSV and disambiguated three more repeated names with different data (Jim Wright, Mel Stottlemyre, Dave Stapleton) using the same (1)/(2) approach.
4. Filtered out players with no recorded batting average from the height versus weight view, since a missing average generally indicates a player who did not bat enough to be meaningful for the analysis.
5. Added calculated fields for BMI, BMI bin (every 0.5), height-to-weight ratio, and a BMI-based color category, then built additional views: batting average versus home runs colored by handedness, height versus weight colored by BMI category and sized by home runs, a histogram of home runs and batting average binned by BMI, and a line view examining height against home runs and batting average.
6. Collected feedback from two reviewers (a coworker and the author's wife) and revised the visualization accordingly, adding the handedness color encoding and the BMI histogram and height/home run line view described above.
7. Assembled the worksheets into two Tableau Story dashboards, Baseball Story_v1 and Baseball Story_Final, and published both to Tableau Public.

Published visualizations:

• Version 1: https://public.tableau.com/profile/robert.hickey#!/vizhome/Tableau_Baseball_Final/BaseballStory_v1
• Final version: https://public.tableau.com/profile/robert.hickey#!/vizhome/Tableau_Baseball_Final_Viz/BaseballStory_Final

Key findings

These findings are drawn directly from the project write-up in Tableau_Project.pdf:

• Batting average versus home runs shows a slight positive correlation; when colored by handedness, there is some evidence that ambidextrous players hit fewer home runs than players who bat right- or left-handed.
• Height versus weight shows the expected positive relationship.
• Coloring height versus weight by BMI category (above or below 25) and sizing by home runs suggested a possible slight positive relationship between BMI and both home runs and batting average.
• Binning the data by BMI (every 0.5) in a histogram made the BMI-to-home-runs relationship clearer, though the relationship between BMI and batting average appeared weaker. Players with a very low BMI (around 19.5) showed a notably high batting average in one bin, but the sample size at that extreme was small enough that this is not treated as conclusive.
• Some taller players had comparatively small home-run markers, prompting a look at height versus home runs and batting average directly, which suggested a possible slight inverse relationship between height and both measures.

Repository structure

• Tableau_Baseball_v1 and Final.twb - Tableau workbook. Contains seven worksheets (Avg v HR, Avg v HR v Handedness, BMI v Avg and HR, BMI v Avg and HR v1, Ht vs Avg HR and Avg Bat, Ht. Vs. Wt., Ht. Vs. Wt. v1) and two story dashboards (Baseball Story_v1, Baseball Story_Final) built from a single data source, baseball_data.
• baseball_data.csv - source dataset used by the workbook, cleaned as described above.
• baseball_data.hyper - Tableau extract generated from the CSV for faster rendering.
• Tableau_Project.pdf - written project reflection covering the summary, design choices, feedback received, and resources consulted.
• README.md - this file.
• LICENSE - project license.

How to run

Open Tableau_Baseball_v1 and Final.twb in Tableau Desktop, or view the published stories directly on Tableau Public using the links above. The workbook was built with Tableau Desktop 2018.1 (file build 20181.18.0615.1128); this is noted for historical context rather than as a pinned requirement, and the file should open in any reasonably current version of Tableau Desktop.

License

Released under the MIT License. See LICENSE for details.