1
Q
What are descriptive statistics used for in a data analytics role?
A
To summarize data in terms of mean, median, interquartile range, standard deviation, or skewness
Descriptive statistics help in understanding how numerical and categorical data are distributed.
2
Q
Why is it important to understand descriptive statistics?
A
Because machine learning algorithms are based on statistical models
Understanding your data profile is critical for selecting the appropriate algorithm.
3
Q
What is the focus of descriptive statistics?
A
Describing useful information about a given data set
It does not require presenting all of the data.
4
Q
Name some Python packages used to calculate descriptive statistics.
A
- Pandas
- NumPy
- SciPy
These packages provide functions to compute various descriptive statistics.
5
Q
What does the mean represent in descriptive statistics?
A
The average value of a dataset
It is calculated by summing all values and dividing by the number of values.
6
Q
What does the median indicate in a dataset?
A
The middle value when the data is ordered
It is less affected by outliers compared to the mean.
7
Q
What is the interquartile range (IQR)?
A
The range between the first quartile (Q1) and the third quartile (Q3)
It measures the spread of the middle 50% of the data.
8
Q
What does standard deviation measure?
A
The amount of variation or dispersion in a dataset
A low standard deviation indicates that data points tend to be close to the mean.
9
Q
What does skewness indicate about a dataset?
A
The asymmetry of the distribution of values
Positive skew indicates a longer tail on the right, while negative skew indicates a longer tail on the left.
10
Q
True or false: Plotting data can help in understanding its summary better.
A
TRUE
Visualizations can reveal patterns and insights that may not be apparent from summary statistics alone.
11
Q
Name the popular descriptive statistics for representing a ‘typical’ value for a dataset.
A
- Average
- Median
- Mode
These measures help summarize the central tendency of the data.
12
Q
Why do we study Descriptive statistics?
A
Because Data Analytics is based on statistical models
Understanding your data profile is critical to select the appropriate test that best fits your data.
13
Q
What does the term distribution mean in data analytics or statistics?
A
A probability distribution
A distribution refers to the spread of the values across a range.
14
Q
What is a variable?
A
Any characteristic, behaviour, category, or number that can be measured or counted
Variables are fundamental in data analysis and statistics.
15
Q
What are the two main types of numerical variables?
A
- Discrete
- Continuous
Numerical variables can be either whole numbers or values within a range.
16
Q
Define a continuous variable.
A
A variable that may contain any value in a range
Examples include spending amounts or time measured in seconds.
17
Q
Define a discrete variable.
A
A variable that has only particular valid values
Examples include shoe sizes or the number of times you went to the market.
18
Q
What are categorical variables?
A
Variables selected from a group of labels
Examples include coin toss outcomes or marital status.
19
Q
What is cardinality in the context of categorical variables?
A
The number of different labels a categorical variable can have
For example, a coin toss has two outcomes: heads or tails.
20
Q
What is the difference between ordinal and nominal variables?
A
- Ordinal: Related to an order (e.g., days of the week)
- Nominal: No specific order (e.g., preferred color)
Ordinal variables have a ranking, while nominal variables do not.
21
Q
True or false: A test result can be encoded as a number, such as 0 for fail and 1 for pass, making it a discrete variable.
A
FALSE
It is a categorical variable that was encoded, not a discrete variable.
22
Q
What is an example of a situation where a categorical variable is encoded as a number?
A
A test result encoded as 0 for fail and 1 for pass
This illustrates how categorical data can be represented numerically.
23
Q
What is a unique ID in the context of data variables?
A
A set of numbers generated for identification purposes
IDs are typically categorical variables, even if they are numeric.
24
Q
What is the purpose of descriptive statistics?
A
Describing and summarising the data
This includes a quantitative approach (numerical summary) and graphical representation (plots).
25
What type of analysis is performed when describing **one variable**?
Univariate analysis
## Footnote
This focuses on a single variable's characteristics.
26
What type of analysis is conducted when studying **two variables** at once?
Bivariate analysis
## Footnote
This examines the relationship between two variables.
27
What is the term for studying **more than two variables**?
Multivariate analysis
## Footnote
This involves analyzing multiple variables simultaneously.
28
Descriptive statistics can be broken down into two major studies: **central tendency** and _______.
Variability
## Footnote
These studies help summarize the dataset's characteristics.
29
Define **population** in the context of data analysis.
All existing elements from a particular variable
## Footnote
Collecting data from a population can be challenging due to its size.
30
What is a **sample**?
A part of a population
## Footnote
A sample ideally preserves the statistical characteristics of the population if its size is statistically significant.
31
What happens to the error when the **sample size** increases?
The error decreases
## Footnote
A larger sample size leads to more accurate inferences about the population.
32
What are **outliers**?
Data points that differ from most of the data
## Footnote
Outliers can indicate errors or new behaviors in the data.
33
What might cause **outliers** in data?
* New natural behavior of the data
* Error in the data collection process
* Human error or bias
* Equipment not being calibrated
## Footnote
Understanding the context of data collection is crucial for interpreting outliers.
34
True or false: Outliers can exist in data analysis.
TRUE
## Footnote
Recognizing the presence of outliers is important for accurate data interpretation.
35
What is the purpose of **central tendency** in data analysis?
To provide figures that summarize the data
## Footnote
Central tendency includes measures such as mean, median, and mode.
36
What are the **measures of variability** mentioned in the text?
* Variance
* Standard Deviation
* Skewness
* Kurtosis
* Percentiles and Quartiles
* Range
## Footnote
These measures help quantify how data is spread around the mean.
37
What does **variance** show?
The spread of the data around the mean
## Footnote
It indicates how far/close the data points are from the mean.
38
What is the symbol for **standard deviation**?
σ
## Footnote
Standard deviation measures the dispersion of the data from the mean in either direction.
39
How can you calculate **standard deviation** in a DataFrame or Series?
* Using the square root of variance
* Using the method .std()
## Footnote
np.sqrt() can be used for the square root calculation.
40
What does **skewness** measure?
The asymmetry of the data
## Footnote
A distribution is symmetric when it looks the same to the left and right of the center point.
41
What happens in a distribution with **positive skewness**?
The tail on the right side is longer
## Footnote
This indicates that there are more extreme values on the higher end.
42
What does **negative skewness** indicate?
The tail of the left side of the distribution is longer
## Footnote
This suggests more extreme values on the lower end.
43
How can you calculate **skewness** in a DataFrame or Series?
Use .skew()
## Footnote
This method provides the skewness value for the data.
44
What does **kurtosis** relate to?
Studying the tails of the distribution
## Footnote
It measures the presence of outliers in the distribution.
45
What does high **kurtosis** indicate?
Significant tails or many outliers
## Footnote
This suggests a distribution with extreme values.
46
What is a **percentile**?
All values below a given percentage
## Footnote
For example, the 50th percentile includes all values below which 50% of observations may be found.
47
What are the three **quartiles** mentioned?
* First quartile (Q1 or 25th percentile)
* Second quartile (Q2 or 50th percentile or median)
* Third quartile (Q3 or 75th percentile)
## Footnote
Quartiles divide the percentiles into four parts.
48
What is the **interquartile range (IQR)**?
The difference between the first and third quartile
## Footnote
It helps understand the range where 50% of the most frequent data is.
49
How can you calculate **quartiles** in a DataFrame or Series?
Use .quantile()
## Footnote
This method provides the quartile values for the data.
50
What is the **range** in statistics?
The difference between the maximum and minimum values
## Footnote
It provides a measure of the spread of the data.
51
How can you find the **minimum value** in your numerical data?
Use .min()
## Footnote
This method retrieves the smallest value in the dataset.
52
What is the purpose of a **statistical test**?
To measure a test statistic that explains the relationship between variables
## Footnote
Statistical tests are essential for analyzing data and drawing conclusions.
53
Name the **types of statistical tests** considered in data analytics.
* Normality tests
* Correlation tests
* Parametric tests
* Non-parametric tests
## Footnote
These tests are grouped based on their objectives.
54
What is the **Shapiro-Wilk test** used for?
To evaluate if data is normally distributed
## Footnote
It is a type of normality test.
55
List three **correlation tests** mentioned.
* Pearson's
* Spearman's
* Chi-Squared Test
## Footnote
These tests evaluate how variables correlate.
56
What are **parametric tests** used for?
To evaluate if data from distinct groups are similar or different
## Footnote
Examples include Student's t-test and ANOVA.
57
Name two **non-parametric tests** mentioned.
* Mann-Whitney U Test
* Wilcoxon Test
## Footnote
Non-parametric tests do not assume a specific data distribution.
58
What does the **Kruskal-Wallis test** evaluate?
If data from distinct groups are similar or different
## Footnote
It is a non-parametric alternative to ANOVA.
59
Why do we need to use **statistical tests**?
* Determine differences or similarities between groups
* Evaluate if a predictor variable has statistical importance to a target variable
## Footnote
Statistical tests provide insights into data relationships.
60
Fill in the blank: A **test statistic** is a number that explains the _______ between your variables in the test.
relationship
## Footnote
Understanding the relationship is crucial for interpreting statistical tests.
61
What does the **Chi-Squared Test** measure?
Significant difference between expected and observed frequencies in categorical variables
## Footnote
It assesses how well observed data fits with expected data.
62
What is the **null hypothesis** in hypothesis testing?
No difference in frequency or proportion of occurrences in each category
## Footnote
It serves as a default position that indicates no effect or no difference.
63
What is the **alternate hypothesis** in hypothesis testing?
There is a difference in frequency or proportion of occurrences in each category
## Footnote
It typically represents the research question being tested.
64
What is the purpose of **hypothesis testing**?
Forming opinions or conclusions from collected data
## Footnote
It involves comparing observed data to expected data.
65
What is the **Significance Level** (alpha)?
Probability of rejecting the null hypothesis when it is true
## Footnote
Commonly set at 5%, indicating a 5 in 100 chance of error.
66
What does a **test statistic** represent?
A number explaining how different the relationship between variables is
## Footnote
It varies in calculation depending on the type of statistical test.
67
What is a **p-value**?
Probability that the null hypothesis is true
## Footnote
A smaller p-value indicates stronger evidence against the null hypothesis.
68
If the **p-value** is lower than the alpha, what can be concluded?
Enough evidence to reject the null hypothesis
## Footnote
This indicates that the observed effect is statistically significant.
69
What does the **Shapiro-Wilk test** assess?
Whether a given data set is normally distributed
## Footnote
The null hypothesis states that the population is normally distributed.
70
What happens if the p-value in the Shapiro-Wilk test is less than the chosen alpha level?
Reject the null hypothesis
## Footnote
This suggests that the data is not normally distributed.
71
What are the two main types of **statistical tests**?
* Parametric tests
* Nonparametric tests
## Footnote
The choice between these tests depends on the normality of the data.
72
When should you use a **parametric test**?
When the data is **normally distributed**
## Footnote
Parametric tests assume that the underlying data follows a normal distribution.
73
When should you use a **nonparametric test**?
When the data is **not normally distributed**
## Footnote
Nonparametric tests do not assume a specific distribution of the data.
74
What is a **t-test** used for?
To test the difference between **two sample means**
## Footnote
It tests if the difference in the means is 0.
75
Who developed the **t-test**?
William Gosset of **Guinness's Brewery**
## Footnote
The t-test is also known as Student's t-test.
76
What is the null hypothesis in a **t-test**?
There are **no significant levels of difference** between the samples
## Footnote
The alternative hypothesis states that there are significant levels of difference.
77
What is a **Paired Student’s t-test** used for?
To test the difference between **two sample parameter values**
## Footnote
The samples should be dependent or paired.
78
What type of samples are required for a **Paired Student’s t-test**?
The samples should be **dependent (paired)**
## Footnote
An example is testing the same group before and after an intervention.
79
What does **ANOVA** stand for?
Analysis of **Variance**
## Footnote
ANOVA compares mean variation between three or more groups.
80
What is required for the data in an **ANOVA** test?
The data should be **normally distributed**
## Footnote
ANOVA tests assume normality in the data distribution.
81
What is an example dataset that could be used in an **ANOVA** test?
Pain threshold levels across different people with varying **hair colors**
## Footnote
This dataset can show mean variations among groups.
82
What is the **Mann-Whitney U Test** used for?
To determine differences between two groups where at least one group is not normally distributed
## Footnote
It is a nonparametric test and requires independent (unpaired) samples.
83
True or false: The **Mann-Whitney U Test** is a parametric test.
FALSE
## Footnote
The Mann-Whitney U Test is a nonparametric test.
84
What type of samples are required for the **Mann-Whitney U Test**?
* Independent samples
* Unpaired samples
## Footnote
This test is used when at least one group is not normally distributed.
85
What is a **Wilcoxon Test** used for?
To analyze paired samples when at least one sample is not normally distributed
## Footnote
It is a non-parametric test and requires dependent (paired) samples.
86
What type of samples are required for the **Wilcoxon Test**?
* Dependent samples
* Paired samples
## Footnote
This test is suitable for matched pairs, such as before and after measurements.
87
Give an example of when to use a **Wilcoxon Test**.
When examining the difference between scores on a test before and after a training intervention
## Footnote
This involves the same group being tested twice.
88
What is the **Kruskal-Wallis test** used for?
To determine differences between three or more groups when at least one distribution is not normally distributed
## Footnote
It is a nonparametric alternative to a one-way ANOVA.
89
True or false: The **Kruskal-Wallis test** is a parametric alternative to a one-way ANOVA.
FALSE
## Footnote
The Kruskal-Wallis test is a nonparametric test.
90
What is a key characteristic of the **Kruskal-Wallis test**?
It is used for three or more groups
## Footnote
It assesses whether there are differences among groups when at least one distribution is not normally distributed.
91
What is **A/B testing**?
A method of comparing two versions of a product to determine which one performs better
## Footnote
Widely used in web development, marketing, and product design.
92
In A/B testing, what are the two versions typically referred to as?
* Version A
* Version B
## Footnote
These versions are shown to different segments of users simultaneously.
93
What is the purpose of **A/B testing**?
To make data-driven decisions that can lead to improvements in user experience and business outcomes
## Footnote
It helps identify which version of a product performs better based on specific metrics.
94
What is an example of an **A/B test** involving background color?
Testing a webpage with:
* Version A: White background
* Version B: Light blue background
## Footnote
The objective is to determine which background color leads to a higher conversion rate.
95
What is the first step in designing an **A/B test**?
Identify the objective clearly
## Footnote
The objective should be specific, measurable, attainable, relevant, and time-bound (SMART).
96
What is a potential objective for an **E-commerce website** in A/B testing?
Increase the conversion rate on product pages
## Footnote
For example, testing whether adding customer reviews increases sales.
97
What is a potential objective for an **email marketing campaign** in A/B testing?
Improve the open rate of a promotional email
## Footnote
For example, testing two different subject lines to see which leads to a higher open rate.
98
What is a potential objective for a **mobile app** in A/B testing?
Enhance user engagement
## Footnote
For example, testing whether adding a daily motivational quote feature increases average time spent on the app.
99
What is a potential objective for **brick-and-mortar retail** in A/B testing?
Increase sales of a product in the store
## Footnote
For example, testing whether showing an ad at the store's entrance increases sales.
100
What is the significance of **randomization** in A/B testing?
Visitors are randomly assigned to different versions to ensure unbiased results
## Footnote
This helps in accurately measuring the performance of each version.
101
What is the importance of **sample size** in A/B testing?
To ensure a large enough sample size to detect a meaningful difference
## Footnote
For example, having 10,000 visitors in each version can provide reliable data.
102
True or false: A/B testing can only be applied to webpages.
FALSE
## Footnote
A/B testing can be applied to various products, including apps and marketing campaigns.
103
What did Google test in their A/B test regarding link colors?
41 different shades of blue for its link colors
## Footnote
The objective was to identify the most effective color for increasing click-through rates.
104
What was the objective of Facebook's A/B test on ad formats?
To increase user interaction with ads
## Footnote
Users were randomly assigned to see one of two ad formats.
105
What did Amazon test in their A/B testing regarding product pages?
Different layouts of the product detail page
## Footnote
The objective was to see which layout led to higher conversion rates.
106
What is the first step after clarifying the **objective** in A/B testing?
Create hypotheses
## Footnote
A hypothesis is a statement that can be tested and will either be supported or rejected by the test results.
107
Define a **hypothesis** in the context of A/B testing.
A statement that can be tested
## Footnote
Each hypothesis should be specific and measurable.
108
What are the two main approaches in A/B testing for hypothesis statements?
* 2-sample cases
* 1-sample cases
## Footnote
A two-sample hypothesis test compares the performance metric of two different versions.
109
What does the **null hypothesis (H-zero)** state in a two-sample A/B test?
No significant difference between conversion rates
## Footnote
Any observed difference is due to random chance.
110
What is the **alternative hypothesis (H1)** in A/B testing?
There is a significant difference between conversion rates
## Footnote
The observed difference is not due to random chance.
111
In a one-sample test, what does the null hypothesis state?
No significant difference between the sample statistic and the known industry standard
## Footnote
Any observed difference is due to random chance.
112
What is the **Minimum Detectable Effect (MDE)**?
The smallest improvement you want to detect
## Footnote
For example, a 5 percent increase in the conversion rate.
113
What is the significance level (alpha) typically set at in A/B testing?
0.05
## Footnote
This means you are willing to accept a 5 percent chance of a false positive.
114
What does **power (1 - Beta)** represent in A/B testing?
The probability of detecting a true effect
## Footnote
Typically set at 0.8, meaning an 80% chance.
115
What is the purpose of **randomisation** in A/B testing?
To randomly assign users to control or treatment groups
## Footnote
This helps eliminate biases and confounding variables.
116
What should be maintained to improve the statistical power of an A/B test?
Equal group sizes
## Footnote
This ensures that the test results are more reliable.
117
What is crucial during the **implementation phase** of A/B testing?
Setting up and launching the A/B test correctly
## Footnote
Any errors can compromise the integrity of the test results.
118
What are **key performance indicators (KPI)** used for in A/B testing?
* Conversion rates
* Click-through rates
* Engagement metrics
* Retention rates
## Footnote
These metrics measure the success of the test.
119
What does **statistical analysis** help determine in A/B testing?
Whether differences between groups are statistically significant
## Footnote
It evaluates if observed differences are due to random chance.
120
What is a common threshold for statistical significance in A/B testing?
p-value less than 0.05
## Footnote
Indicates that the results are statistically significant.
121
What does a **confidence interval** provide in A/B testing?
A range of values for the true effect of the treatment
## Footnote
A typical value might be a 95% confidence interval.
122
What should be assessed when interpreting A/B test results?
* Magnitude of the effect
* Business impact
* Cost and feasibility
* Long-term implications
## Footnote
This ensures informed decisions that drive business outcomes.
123
True or false: Focusing solely on **statistical significance** is sufficient for interpreting A/B test results.
FALSE
## Footnote
Always assess practical implications and real-world impact.
124
What common pitfall involves overlooking external factors in A/B testing?
Ignoring external factors
## Footnote
These can influence test results, such as seasonality or market trends.
125
What does **insufficient sample size** lead to in A/B testing?
Unreliable results
## Footnote
Ensure the sample size is large enough to detect a meaningful effect.
126
What is the risk of **cherry-picking results** in A/B testing?
Selectively reporting favourable results
## Footnote
Report all results transparently and conduct thorough analysis.
127
What should be evaluated to avoid a **short-term focus** in A/B testing?
Long-term impact of changes
## Footnote
Conduct follow-up tests if necessary.
128
Fill in the blank: A simple business problem might be: The webpage ‘contact us’ form is too long, so the hypothesis is that reducing the form inputs from 20 to 10 will _______.
increase completed forms
## Footnote
This can be evaluated by A/B testing the two forms.
129
What statistical method might be used to evaluate a hypothesis about tool maintenance in a factory?
Regression model
## Footnote
This model predicts the best time for maintenance based on data collected.
Data Analytics
flashcards
Decks in class (27)
# Cards
-
Productivity & Growth Mindset82
-
What Is Data Analytics69
-
Data Cleansing26
-
Data Storage43
-
Data Transformation30
-
Data Management43
-
ETL & SAP31
-
Data Governance25
-
Data Modelling22
-
Business Requirements55
-
Project Coordination20
-
Analysing Problems & Developing Solutions46
-
TeamWork35
-
Stakeholder Management & Emotional Intelligence80
-
Exploratory Data Analytics Techniques31
-
Design Requirements67
-
Probability and Statistics129
-
Python- Libraries-Streamlit26
-
ML134
-
SQL11
-
Power BI30
-
Image & Text Analysis28
-
Neural-network Algorithms.67
-
ML Model Analysis96
-
ML OPs70
-
Implement AI Model112
-
Ethics113
