Data Training Courses

Home | Power BI | Excel | Python | SQL | Generative AI | Visualising Data | Analysing Data

Iris Dataset

The Iris dataset is a classic dataset from the 1930s. The data contains 150 rows (observations) of irises; 50 from each of 3 species: setosa, versicolor and virginica. Here is an example of each species

setosa

versicolor

virginica

Data Description and Links

The dataset has 6 columns:

• observation – a unique number
• petal length (in cm)
• petal width
• sepal length
• sepal width
• species The data is in CSV format in a public file. The URL is https://zomalextrainingstorage.blob.core.windows.net/datasets/misc/iris.csv.

The data looks like this (not all rows shown)

Suggested Approach

1. Import the iris data CSV file
2. Build a scatter plot of petal length vs petal width and use colour to differentiate between species
3. Explore different formatting options to make a better chart; e.g. colour, shape, fill point, axes start and end values

Power BI lab

If analysing the data with Power BI, here are some helpful hints

• While in Query Editor check the Column Quality and Column distribution boxes in the View pane. What does this tell you about the data?
• Set the default summarisation of the Observation column to “Don’t summarize”
• Create a measure named to count the number of Irises. e.g., Number of Irises = COUNTROWS(irisdata)
• Use ColorBrewer to pick a palette of three qualitative colours and replace Power BI’s default colours with these
• Choose helpful labels for both axes and a good title for the chart
• Add labelled images of irises to the page or even as a background to the chart?
• Add an overall trend line or a trend line for each species. Is this helpful?