In a situation where the initial 'questioning' phase did not yield actionable insights, what might be the next step in the EDA process?
- Jump to the concluding phase to draw insights
- Proceed to the exploring phase without adjustment
- Revisit the questioning phase to refine or develop new questions
- Skip to the communication phase
If the initial 'questioning' phase does not yield actionable insights, it is necessary to revisit the questioning phase to refine or develop new questions. The questions set the direction of the analysis and are crucial for subsequent steps. If the questions are not well defined or not actionable, it could lead to an ineffective analysis.
_____ data is numerical in nature and can be ordered or measured.
- Nominal
- Ordinal
- Qualitative
- Quantitative
Quantitative data is numerical, measurable, and can be used with mathematical operations.
How does the 'subplot' function in Matplotlib differ from 'FacetGrid' in Seaborn?
- FacetGrid allows the creation of multi-plot grids based on row and column-wise grouping of the data
- FacetGrid supports interactive plotting
- Subplot can create only single plots
- Subplot does not allow the sharing of axes
The 'subplot' function in Matplotlib is used for creating sub-plots in a single figure, but it does not allow easy creation of grids of plots based on categorical variables. On the other hand, 'FacetGrid' in Seaborn allows the creation of complex multi-plot grids based on row and column-wise grouping of the data.
Under what conditions would the median be a better measure of central tendency than the mean?
- When the data has an even number of observations
- When the data has outliers or is skewed
- When the data is normally distributed
- When the data is uniformly distributed
The median would be a better measure of central tendency than the mean when the data has outliers or is skewed. In these cases, the mean can be heavily influenced by the extreme values, while the median, being the middle value, remains more robust and representative of the central location of the data.
How is model-based method different from the other two imputation methods?
- It deletes missing data
- It estimates missing values based on a statistical model
- It is not different from the others
- It uses the mode value for imputation
The model-based method is different from the other imputation methods as it estimates missing values based on a statistical model. This method assumes a specific statistical model (like a linear regression, logistic regression, etc.) that generates the data, and missing values are filled in based on this model.
In what scenario would a modified Z-score be beneficial to use for outlier detection?
- When data is bimodal
- When data is normally distributed
- When data is skewed or has outliers
- When data is uniformly distributed
A modified Z-score is beneficial to use for outlier detection when data is skewed or has outliers, as it is more robust to outliers than the traditional Z-score.
You have a dataset with a large number of missing values. What strategies can you use to depict this in your data visualization?
- Ignore the missing values, because they can't be visualized
- Only include complete cases in the visualization
- Replace all missing values with the mean
- Use a different color or pattern to indicate missing values
Missing values can be indicated in data visualizations using a different color or pattern. This strategy allows viewers to see where data is missing, which can be informative in itself. Ignoring or inaccurately replacing missing values can lead to misleading visualizations.
Which type of missing data is completely random and does not depend on any observed or unobserved data?
- MAR
- MCAR
- NMAR
- nan
MCAR (Missing Completely At Random) indicates that the missingness of data is completely random and does not depend on any observed or unobserved data.
During your EDA process, you identify several outliers in your dataset. How does this finding impact your subsequent steps in data analysis?
- You may need to collect more data
- You may need to ignore these outliers as they are anomalies
- You might consider robust methods or outlier treatment methods for your analysis
- You might decide to use a different dataset
Identifying outliers during the EDA process would influence the subsequent steps in data analysis. The outliers could indicate errors, but they could also be true data points. Depending on the context, you might need to investigate the reasons for their presence, treat them appropriately (for example, using robust statistical methods, data transformations, or outlier removal), or revise your analysis techniques to accommodate them.
In a scatter plot, outliers often appear as points that are far removed from the ___________.
- axes
- main concentration of data
- origin
- trend line
In a scatter plot, outliers are often represented as points that are far removed from the main concentration of data.