What aspects should be considered to improve the readability of a graph?
- All of the mentioned
- The amount of data displayed
- The color scheme
- The scale and labels
Improving the readability of a graph involves considering several aspects, including the color scheme (which should be clear and not misleading), the scale and labels (which should be appropriate and informative), and the amount of data displayed (too much data can overwhelm the audience and obscure the main message).
You are given a dataset with a significant amount of outliers. Which scaling method would be most suitable and why?
- , outliers should always be removed
- Min-Max scaling because it scales all values between 0 and 1
- Robust scaling because it is not affected by outliers
- Z-score standardization because it reduces skewness
Robust scaling would be the most suitable as it uses the median and the interquartile range, which are not sensitive to outliers, for scaling the data. Other methods like Min-Max and Z-score are affected by the presence of outliers.
Under what circumstances might 'removal' of outliers lead to biased results?
- When outliers are a result of data duplication
- When outliers are due to data collection errors
- When outliers are extreme but legitimate data points
- When outliers do not significantly impact the analysis
Removing outliers can lead to biased results when the outliers are extreme but legitimate data points, as they could represent important aspects of the phenomenon being studied.
You have applied mean imputation to a dataset where values are missing not at random. What kind of bias might you have unintentionally introduced, and why?
- Confirmation bias
- Overfitting bias
- Selection bias
- Underfitting bias
If you have applied mean imputation to a dataset where values are missing not at random, you might have unintentionally introduced selection bias. This is because mean imputation could lead to an underestimation of the variability in the data and potentially introduce a systematic bias, as it doesn't consider the reasons behind the missingness.
What is the main purpose of data normalization in machine learning?
- To ensure numerical stability and bring features to a comparable scale
- To increase the accuracy of the model
- To increase the size of the dataset
- To reduce the computation time
Data normalization is a technique often applied as part of data preprocessing in machine learning to bring features to a comparable scale. This is done to ensure numerical stability during the computation and to prevent certain features from dominating others due to their scale.
What is the importance of the 'explore' step in the EDA process?
- To analyze and investigate the data
- To clean and transform data
- To communicate the results
- To pose initial questions
The 'explore' step in the EDA process is crucial as it involves the analysis and investigation of the cleaned and transformed data, using statistical techniques and visualization methods. This stage helps uncover patterns, trends, relationships, and anomalies in the data, and aids in forming or refining hypotheses.
What is the importance of understanding data distributions in Exploratory Data Analysis?
- All of the above
- It helps in identifying the right statistical tests to apply
- It helps in spotting outliers and anomalies
- It helps in understanding the underlying structure of the data
Understanding data distributions is fundamental in Exploratory Data Analysis. It aids in understanding the structure of data, identifying outliers, formulating hypotheses, and selecting appropriate statistical tests.
Imagine you are working with a data set that includes survey responses on a 1-5 scale (1=Very Unsatisfied, 5=Very Satisfied). How would you classify this data type?
- Continuous data
- Interval data
- Nominal data
- Ordinal data
This type of data is ordinal because the ratings exist on an arbitrary scale where the rank order (1-5) is significant, but the precise numerical differences between the scale values are not.
You have a large dataset where removing the outliers would lead to loss of significant data. What method would you recommend for outlier handling?
- Binning
- Removal
- Transformation
- nan
If the dataset is large and removing outliers would lead to a significant loss of data, binning could be a suitable method. In binning, the outliers are not removed but rather they are replaced with summary statistics like mean, median, etc.
Consider you are dealing with a dataset with zero skewness but high kurtosis. How would this shape the data distribution and affect your analysis?
- The data distribution would be negatively skewed with a wider spread.
- The data distribution would be perfectly symmetrical with a narrower spread and potential outliers.
- The data distribution would be perfectly symmetrical with a wider spread.
- The data distribution would be positively skewed with a narrower spread.
Zero skewness means the distribution is symmetrical, and high kurtosis means the distribution is leptokurtic with a sharp peak and fatter tails. Therefore, the data distribution will be symmetrical but with a potential for outliers. This may affect the results of statistical tests or models that assume normality, as extreme values could have a disproportionate effect on the results.