Machine Learning_review_day3
World Happiness Report
https://www.kaggle.com/datasets/ajaypalsinghlo/world-happiness-report-2021
Visualize as a responsive graphs (treemap(), sunburst(), choropleth())
treemap: data_frame=dataframe object, path=[parent column, child column], values=column property, color=column property
sunburst: data_frame=dataframe object, path=[parent column, child column], values=column property, color=column property
choropleth: data_frame=dataframe object, locations='column names', locationmode='country names', color='column names'
< By Year >
heatmap(im.show): px.imshow(dataframe object.corr( ), text_auto = True), text_auto=True: real value, correlation visualization
scattermatricx: px.scatter_matrix(dataframe object, dimensions=['property name'], color='property name'
==>
- The degree to which the value of another attribute (Ladder score) changes together with the value of the attribute (Logged GDP per capita) is strong, with a high positive correlation with a correlation coefficient of about 0.79.
- The degree to which the value of another attribute (Ladder score) changes together with the value of the attribute (Freedom to make life choices) is moderate, with a moderate positive correlation with a correlation coefficient of about 0.61.
- Since the correlation coefficient is about -0.02, the value of other attributes (Ladder score) hardly changes according to the value of the attribute (Generosity).
- It is a normal negative correlation with a correlation coefficient of about -0.42, and the extent to which the value of another attribute (Ladder score) changes together with the value of the attribute (Perceptions of corruption) is normal.
<Linear Regression Model>
CO2 Emissions Forecast
https://www.kaggle.com/datasets/debajyotipodder/co2-emission-by-vehicles