Analysing Cat Hunting Behaviour

Proposal

Dataset

tag_id animal_id animal_taxon deploy_on_date deploy_off_date hunt prey_p_month animal_reproductive_condition animal_sex hrs_indoors n_cats food_dry food_wet food_other study_site age_years
Tommy-Tag Tommy Felis catus 2017-06-03 01:02:09 2017-06-10 02:10:52 TRUE 12.5 Neutered m 12.5 2 TRUE TRUE FALSE UK 11
Athena Athena Felis catus 2017-06-24 01:02:13 2017-06-30 23:59:32 TRUE 3.0 Spayed f 7.5 2 TRUE TRUE FALSE UK 3
tag_id event_id visible timestamp location_long location_lat ground_speed height_above_ellipsoid algorithm_marked_outlier manually_marked_outlier study_name
Ares 3395610551 TRUE 2017-06-24 01:03:57 -5.113851 50.17032 684 154.67 FALSE FALSE Pet Cats United Kingdom
Ares 3395610552 TRUE 2017-06-24 01:11:20 -5.113851 50.17032 936 154.67 FALSE FALSE Pet Cats United Kingdom

Description of data

The Pet Cats data set comes from Movebank for Animal Tracking Data, collected from volunteers using GPS sensors on their pet cats around UK, US, Australia and New Zealand. The UK dataset contains two subsets called cats_uk and cats_uk_reference. So does the US dataset.

cats_uk contains 18215 observations of 11 variables providing data such as ground speed, geographic longitude and latitude estimated by the sensor, time-stamp for sensor measurements and an event id. These are all numerical variables.

cats_uk_reference contains 101 observations of 16 variables and primarily describes the number of cats living in a house, whether they are allowed to hunt, their prey per month and age. These are numerical data. Also, the data contains the nature of the food that the cats eat, in the form of a logical variable. reproductive condition and animal sex are categorical variables.

Reason of choosing the dataset

We selected this dataset because we can gain insights into the habits of cats by finding a relationship between hunting behavior, prey consumption and hours spent indoors. We all found the dataset interesting, as cats have fascinating characteristics. There seems to be potential relations between hunting capacities, reproductive condition, food habits, age. Representing them through visualizations would provide a great learning curve as the data is varied.

Questions

1) Are the cats which live together better at hunting than those which live alone?

2) How does the reproductive status of cats influence the hunting pattern?

Analysis plan

Question 1

  • The first question is about studying whether the cats who live in group have the ability to hunt better than the cats who live alone. To analyze the above question we will consider some features from the dataset: n_cats (number of cats living in the house), prey_p_month (prey caught per month), age_years (cat’s age). We have categorized the cats into three distinct age groups: kittens (0 to 5 years), young cats (5 to 10 years), and mature cats (above 10 years). The x-axis displays the number of cat data, while the y-axis represents the prey caught per month. The plot is divided into three segments based on age groups. We believe that segmenting the jitter plot by age group is a valuable approach for assessing the hunting capabilities of cats in relation to their age.

  • We will then use a jitter plot to visualize the relationships between these variables effectively. In our plot the values are encoded as:

    • number of cats on the x-axis
    • prey caught per month on the y-axis,
    • points on the plot will be grouped by age.

  • A line plot can also be a suitable plot here, but the information gain will be apparent only upon actual experiments with data and visualization.

Question 2

  • The second question asks whether there is a difference in the hunting pattern based on the reproductive condition of the cat. The dataset contains animal_reproductive_condition that specifies whether the animal was neutered or not. We also have prey_p_month and age that reflect the hunting pattern with respect to age.

  • We are planning to go with a Ridgeline Plot or a Hexbin Plot that will effectively compare the three variables together.

    • Two of the axes of either graph will contain the numeric variables prey_p_month and hrs_indoors.

    • The third categorical variable will help further to categorize (differentiate) the groups in the plot.

  • This visualization will be finalized based on the amount of information retrieval from either if the charts mentioned in the above point.

Source:

Information about the dataset description is derived from the following Tidy Tuesday page: Pet Cats.