Exploratory data analysis of Irises

1. We examine individual columns

1.1 Let's check how many individual irises reach the specific measure values of their structural elements.

We observe that the most irises achieve:

• a sepal length of 5cm,
• a sepal width of 3cm,
• a petal length of about 1.1cm,
• a petal width of about 0.2cm.

1.2 We will now group the irises by all available species and check the individual means, medians, minimum and maximum values of the elements of their structure.

1.2.1 Graphs of average and median sizes of construction elements for individual species

We can observe that:

• The largest average length of the petal is achieved by Iris-virginica, and the median value of petal length is close to the average in each species, so we expect that there are few irises with significantly shorter or longer petal lengths.
• The largest average width of the petal is achieved by Iris-setosa, and the median value of petal width is close to the average in each species, so we expect there are few irises with significantly shorter or longer petal widths.
• On the other hand, significantly the smallest average length of the sepal is achieved by Iris-setosa, and the median value of sepal length is close to the average in each species, so we expect there are few irises with significantly shorter or longer sepal lengths.

1.2.2 Bar chart of minimum and maximum values for individual species

The above chart shows something we have already observed in the bar chart of the average values of flower petal dimensions, namely, we see that Iris-setosa has significantly smaller minimum petal sizes than occurs in other species.

1.5 We will then discuss the histograms of all values for each species.

1.5.1 Histogram of flower structure for Iris-setosa

1.5.2 Histograms of flower structure for Iris-versicolor

1.5.3 Histograms of flower structure for Iris-virginica

Comparing the histograms, we observe that:

• The species Iris-setosa and Iris-versicolor have the highest number of individuals with a sepal length of from 5 to 5.5 cm, while Iris-virginica has the most individuals reaching 6.2 cm
• For the 3 species, the sepal widths that have the most representatives of each species are similar and amount to about 3 cm
• However, the petal lengths achieved by different individuals vary significantly, with Iris-setosa standing out the most, where the majority of representatives reach 1.5 cm, for Iris-versicolor it is 4.5 cm, and for Iris-virginica it is 5.2 cm.

2. Searching for correlations in the data

2.1 We check the correlations between the individual components of the Iris-setosa structure.

We observe a strong correlation in the species Iris-setosa between sepal length and sepal width, meaning that the larger the sepal lengths are, the more flowers of this species we observe with larger sepal widths.

2.2 We check the correlations between the various structural elements of Iris-versicolor

We observe a strong correlation in the species Iris-versicolor between petal length and sepal length, meaning that the larger the petal lengths, the more flowers of this species we observe with larger sepal lengths.

We observe a strong correlation in the species Iris-versicolor between petal length and sepal width, meaning that the larger the petal lengths, the more flowers of this species we observe with larger sepal widths.

2.3 We check the correlations between the individual components of the Iris-virginica structure.

We observe a very strong correlation in the species Iris-virginica between sepal length and petal length, meaning that the larger the sepal lengths are, the more flowers** of this species we observe with larger petal lengths.

3. Searching for outliers

3.1 Box plot of Iris-setosa

We see that the species Iris-setosa has flowers with petal lengths and widths that deviate from the norm of the flowers of this species studied.

3.2 Box plot of Iris-versicolor

We see that the species Iris-versicolor has flowers that have petal lengths deviating from the norm of the studied flowers of this species.

3.3 Boxplot of Iris-virginica

We see that the species Iris-virginica has flowers or flower(s) that have cup lengths that deviate significantly from the norm for the studied flowers of this species. It can also be observed that there are flowers with cup widths that are smaller or larger than the norm for this species.

4. Summary

4.1 Summary of the analysis of columns

• The largest average length of the sepals, which measures 6.2 cm, is achieved by the flowers of the species Iris-virginica, which is slightly more than the flowers of Iris-setosa or Iris-versicolor, where these flowers have an average sepal length of 5 to 5.5 cm.

• The largest average width of the sepals is achieved by the flowers of the species Iris-setosa. * Iris-setosa also has the significantly smallest average petal length, namely 1.5 cm, compared to the other two species, where for Iris versicolor it is 4.5 cm, and for Iris-virginica it is 5.2 cm.

• In representatives of Iris-virginica, most individuals have petal lengths of from 4.5 to 7, while in Iris-versicolor it is from 3 to 5 cm.

4.2 Summary of Correlation Analysis

• In the species Iris-setosa, the greater the lengths of the sepals, the more flowers of this species we observe with larger widths of the sepals.

• In the species Iris-versicolor, the greater the lengths of the petals, the more flowers of this species we observe with larger lengths of the sepals.

• In the flowers of Iris-versicolor, we observed that the greater the lengths of the petals, the more flowers of this species we observe with larger widths of the petals.

• In the species Iris-virginica, the greater the lengths of the sepals, the more flowers of this species we observe with larger lengths of the petals.

[NbConvertApp] Converting notebook iris.ipynb. to html
[NbConvertApp] WARNING | Alternative text is missing on 14 image(s).
[NbConvertApp] Writing 879102 bytes to iris.html