My first analysis project: cleaning a public CSV, building an Excel dashboard, and exploring breed, size, and obedience measures with Excel, MySQL, and Python.
Original Excel dashboard from my first personal data project. IQ is the original dashboard label; the analyzed fields describe working/obedience performance.
01 / THE QUESTION
The question
Does breed, height, or weight relate to the dataset’s dog IQ measure? The available fields describe working/obedience performance, so I used those as the focus of the analysis.
02 / THE APPROACH
The approach
Downloaded a practice CSV from a public-dataset website and cleaned the data in Excel.
Built an Excel dashboard to explore breed groups, working/obedience classifications, and size categories.
Loaded the data into MySQL and inspected the records with SQL.
Used Excel regression and single-factor ANOVA to examine height and compare breed-group means.
Explored weight with a Python regression plot and reviewed relationships with Pandas correlation heatmaps.
03 / KEY INSIGHTS
Key insights
The height regression showed a weak fit (R² ≈ 0.013; p ≈ 0.202), and the breed-group ANOVA did not detect a difference in means at the 0.05 level (p ≈ 0.316). The weight plot was nearly flat. These outputs did not support a clear relationship in the analyzed samples; they do not prove that breed or size can never relate to canine abilities.
04 / OUTCOME & IMPACT
Outcome & impact
Completed my first analysis workflow across Excel, MySQL, and Python, producing a dashboard, regression output, an ANOVA comparison, and correlation heatmaps. The project helped me practice moving from a question to cleaned data, visual exploration, and a carefully qualified conclusion.
THE PROJECT IN DETAIL
Project walkthrough
SLIDE 01
My first question, my first dashboard
For my first data project, I downloaded a CSV from a public-dataset website and asked whether breed, height, or weight was related to the dataset’s dog “IQ” measure. I cleaned the data in Excel and built a dashboard to explore breed groups, classifications, and size categories.
The original dashboard uses the term IQ, but the screenshots show working/obedience classifications and an Obey field. I treat those as the measures in this project, not as a complete measure of canine intelligence.
SLIDE 02
Bringing the data into MySQL
I also loaded the data into MySQL to practice working with it in a relational database. The screenshot shows the dogs schema with akc and iq tables, and a SELECT * FROM dogs.iq query returning 164 rows.
This gave me another way to inspect breed names, classifications, obedience values, and repetition ranges. Some displayed records contain NULL values, an important consideration when preparing data for analysis.
SLIDE 03
Testing the relationship with height
I used Excel regression output and a scatter plot to examine height alongside the obedience measure. The saved output reports 127 observations, R² of approximately 0.013, and a height-coefficient p-value of approximately 0.202.
In this fitted model, height explained about 1.3% of the variation in the response. The result did not provide evidence of a linear association at the 0.05 significance level; it does not prove that no relationship could exist.
SLIDE 04
Comparing breed groups with ANOVA
Because breed group is a category, I used a single-factor ANOVA to compare the average measure across Herding, Hound, Non-Sporting, Sporting, Terrier, Toy, and Working groups.
The output contains 140 observations across seven groups, with F approximately 1.189 and p approximately 0.316. At the 0.05 level, this analysis did not detect a difference in group means. That is not a claim that all breeds have identical abilities.
SLIDE 05
Exploring weight in Python
I continued the exploration in Python using Pandas and Seaborn. This plot compares weight_average with Obey, showing individual observations, a fitted line, and an uncertainty band.
The fitted line is nearly flat and slopes slightly downward, offering little visual support for a strong linear relationship in this sample. The screenshot alone does not establish statistical significance.
SLIDE 06
Reviewing relationships with heatmaps
I used correlation heatmaps to review relationships among the available numerical fields and complement the Excel analysis. Together, the saved outputs supported my takeaway that this dataset did not show a clear relationship between the obedience measure and dog size, or a detectable difference across the breed groups tested.
This first project also gave me useful lessons about interpretation: category codes should not be treated as meaningful continuous measurements, and obedience is narrower than intelligence. The SQL query, height regression, and ANOVA show different record counts, so their samples should be checked before drawing broader conclusions.
Walkthrough reconstructed from Gerry Taylor II’s summary and six original screenshots. Reported statistics are transcribed from the saved outputs, not recalculated. The original CSV and dataset-source link were not supplied.
