Statistics
196 bites tagged Statistics — interview questions with model answers, and 60-second explainers.
Missing Data Imputation: Filling in the Blanks
Instead of deleting rows with missing values, imputation makes an educated guess to fill the blanks, preserving your sample size. It's used in survey analysis or time-series data where dropping records would introduce bias.
Differential Privacy: Anonymize Data with Math
Differential Privacy adds mathematical noise to data queries, making it impossible to know if one person's data is included. Tech giants use it to learn from user behavior without seeing individual activity.
Causal DAGs: A Map for Cause and Effect
A Causal DAG is a map of your assumptions about what causes what. It helps you spot hidden "confounder" variables that create misleading correlations. Use it before an analysis to decide which variables to control for, ensuring you measure a true effect.
Propensity Score Matching: Simulating a Randomized Trial
Propensity Score Matching creates a 'fair' comparison from observational data, mimicking a randomized trial. It's used to estimate a treatment's effect by matching treated individuals with similar untreated ones.
SUTVA: The Assumption That Your Treatment Isn't Leaking
SUTVA assumes your treatment on one person doesn't spill over to affect others, and that the treatment is consistent for all. It's a key assumption for A/B tests, but is violated when one person's vaccine protects their unvaccinated neighbor.
Statistical Power: Is Your Test Strong Enough to Find a Real Effect?
Statistical power is your experiment's ability to detect a real effect. A low-power test is like a fishing net with holes too big—the fish swims right through. This is critical for A/B tests. The footgun is concluding "no effect" from a weak test.
A/A Testing: Sanity-Checking Your Experiments
A/A testing is a fire drill for your A/B testing system. You run two identical versions of a page to ensure your tools are working correctly before a real experiment. The biggest footgun is panicking at a false positive; 1 in 20 tests will show.
Randomization: Defeating Bias with Chance
Randomization uses chance to assign subjects to groups, isolating the effect you're testing from hidden variables. It's the foundation of A/B tests and clinical trials. The footgun is confusing 'haphazard' selection with true, unbiased randomization.
Regularization: Penalizing Complexity to Prevent Overfitting
Regularization is a complexity tax on a machine learning model, forcing it to favor simpler patterns over memorizing training data. It's used to prevent overfitting in models like neural networks, improving their performance on new, unseen data.
Naive Bayes: Fast Classification by Assuming Independence
Naive Bayes classifies data by assuming its features are unrelated, like judging a fruit's type by color and shape independently. This makes it fast for tasks like spam filtering or real-time predictions. Its core 'naive' assumption is almost always wrong.
Cross-Validation: Don't Test on Your Training Data
Cross-validation stops a model from 'cheating' by testing it on unseen data. It repeatedly splits your dataset into training and testing portions to simulate real-world performance.
Logistic Regression: Predicting a Yes or No Outcome
Logistic regression predicts a binary outcome (yes/no, 1/0) by calculating the probability of an event. It's used for classifying things like spam email or fraudulent transactions. The footgun is mistaking its output for certainty instead of a probability.
Q-Q Plot: Visually Test if Data Fits a Distribution
A Q-Q plot visually checks if your data fits a theoretical distribution (like normal). If the points form a straight line, it's a match. Use it to validate model assumptions.
Kernel Density Estimation: Beyond the Histogram
A KDE plot is a smoothed histogram, revealing a variable's distribution by summing 'bumps' over each data point instead of using rigid bins. It's great for data exploration, but the 'smoothness' (bandwidth) choice can drastically alter the plot's story.
Violin Plots: Box Plots with a Sense of Shape
A violin plot is a box plot that also shows the data's shape. It reveals multi-modal distributions or skews that a simple box plot would hide. Use it to compare distributions across categories, like server response times.
Pair Plot: See All Your Data's Relationships at Once
A pair plot is a matrix of charts showing every pairwise relationship in a dataset. Use it in exploratory data analysis to spot correlations and distributions at a glance. The footgun: it becomes unreadably large and slow with more than ~10 variables.
Bar Chart: Compare Categories at a Glance
A bar chart is a lineup of rulers for different categories, making it easy to compare values. It's perfect for showing "how much" for discrete items, like monthly sales or user counts by country. The footgun: don't use it for continuous data—use a line chart.
Box Plot: See a Distribution's Shape at a Glance
A box plot is a cheat sheet for a dataset's distribution, showing the median, the middle 50% of data (the box), and outliers. Use it to quickly compare distributions, like server response times across regions. The footgun: the line is the median, not the mean.
Scatter Plot: Visualizing Relationships Between Two Variables
A scatter plot visualizes the relationship between two variables as a cloud of data points. It's used to quickly spot trends, like whether increased study time correlates with higher test scores. The key footgun: correlation is not causation.
Histogram: See the Shape in Your Data
A histogram turns a raw list of numbers into a picture of its distribution, like sorting coins into stacks. It's used to see where data clusters and how it's spread out. The main footgun: the story your histogram tells depends entirely on the bin size.
Box-Cox Transformation: Forcing Skewed Data to Look Normal
The Box-Cox transformation is a statistical lens that reshapes skewed data to better resemble a normal distribution. It helps meet the assumptions of models like linear regression, but it only works on positive data and complicates direct interpretation of…
Hypothesis Testing: A Courtroom for Your Data
Hypothesis testing is a courtroom trial for a claim. You assume 'no effect' (the null hypothesis) and see if your data is strong enough to reject it. The footgun is misreading the p-value: it only measures evidence against the null, not for your alternative.
Linear Regression: Finding the Line of Best Fit
Linear regression draws the 'line of best fit' through your data to predict outcomes. It's used to estimate continuous values, like forecasting sales based on ad spend or predicting a house's price from its size. The main footgun: correlation is not causation.
Bayes' Theorem: Updating Beliefs with Evidence
Bayes' Theorem updates your belief in a cause after seeing new evidence. It's used in medical diagnostics to interpret test results and in spam filters. The common footgun is ignoring the base rate—how likely the cause was *before* the evidence appeared.
Get Statistics bites daily.
Five a day, five minutes, offline. With quizzes so it sticks.
Open testing — you’ll join as an early tester.