2.4 KiB
21 Setembro 2023 - #DAA
-
Types of Data
-Numerical
- Discrete Data ({1,2,3,...})
- Continuous Data ([1, +$\infty$])
-
Categorical (binary, languages, ...)
-
Ordinal (ratings of 1 to 5)
*not-scaled **scaled
[!example]- Qual o tipo de dado que representa a quantidade de gasóleo? Numérico: continuous data
[!example]- Qual o tipo de dado que representa a nacionalidade? Categórico
[!example]- Qual o tipo de dado que representa a idade? Numérico: discreto
-
Mean, Median & Mode
-
#+BEGIN_TIP A mean in math is the average of a data set, found by adding all numbers together and then dividing the sum of the numbers by the number of numbers #+END_TIP
-
[!hint]+ A mean in math is the average of a data set, found by adding all numbers together and then dividing the sum of the numbers by the number of numbers.
-
Standard Deviation & Variance
-
Probability Density functions
-
Percentiles
There are 3 important percentiles:
-
50% - median
-
25% - 1st percentile
-
75% - 3rd percentile
[!note]+ These 3 percentiles allow the creation of box plot graphs. These specific graphs allow the discovery and presentation of outliers.
-
Covariance & Correlation
[!hint]+ Covariance measures the direction of a relationship between two variables, while correlation measures the strength of that relationship.
Covariance is hard to interpret, thus correlation is used instead. In a dataset, correlations >0.5 are considerable.
[!caution] Correlation does not mean causation!
-
Practical session of the class: miniconda
IDEs: PyCharm, VS Code, (Jupyter - not recommended)
Depois de instalar o miniconda, correr os seguintes comandos:
-
conda create --name daaEnv python=3.10
-
conda activate daaEnv
-
python --version
-
conda install pandas
-
conda install xlrd
-
conda install xlwt
-
conda install matplotlib
-
conda install seaborn
-
conda install scikit-learn
-
conda install jupyterlab
-
conda list
-
Resource links