Gallery
Discover apps people built on Collabby by chatting. Run them, then remix to make them yours.
Gradient Descent and Loss Functions
Roll a ball downhill, opposite the slope, to find the minimum—and see how a high learning rate makes it bounce away.
Sampling and Bias
Change how you sample to see where bias comes from, and discover which biases remain even with larger samples.
Box Plots and Data Spread
See how one distant point pulls the mean, then use quartiles to make a box plot and spot outliers.
Binomial Distribution & Expected Value
Move the probability bars to see expected value as the balance point, then change n and p to reshape the binomial distribution.
Scatterplots & Correlation
Read the relationship between two quantities from how the points scatter, fit a trend line by eye, and see why predictions hold only inside the data's range.
Least-Squares Regression & Residuals
Drag the line to find where the sum of squared residuals is smallest, then vary the spread to see how correlation r behaves and why extrapolation is risky.
Significance Tests & p-values
Simulate outcomes under the null hypothesis to estimate the p-value, then shift the significance level α to find where the decision flips.
Sample Means and Confidence Intervals
Draw samples repeatedly to watch the sampling distribution of the mean narrow, and count how often the confidence interval captures the population mean.
Normal Distribution and Standardization
Adjust the mean and standard deviation to shape a normal curve, confirm the 68–95–99.7 rule, and standardize with z to read probabilities.
Anchoring Effect
Change the anchor number you see before each estimate and find out how far apart the answers drift after large numbers versus small ones. 🍴 Remix it to swap the six estimation questions for ones about your own school.
Regression to the Mean
Adjust luck's share in a score and see why the top scorers' average slides back toward the middle on a second test, though no one's skill changed.
Simpson's Paradox
Shift the patient mix to see Simpson's paradox: the hospital ahead in every subgroup falls behind in the pooled total.
Conditional Probability and Base Rate
Adjust the rarity and the test's accuracy to see that the base rate, not the test, decides how likely a positive result is real.
Loss Aversion
Compare gain and loss options with identical expected values, repeat the gamble, and reword the same fact to see that the phrasing, not the number, decides.
Confirmation Bias
Feed in three numbers to get a yes or no, and see that tests aimed at a «no», not confirmations, are what pin down the hidden rule.
Survivorship Bias
Adjust where bullets hit and how deadly each part is to see how the planes that never returned distort the numbers.
Gambler's Fallacy
Count only the tosses that follow a run of heads and watch the share settle at one half while the gap between the counts keeps widening.
Mean and Median
Move seven friends' allowances to see why one huge value sends the mean soaring while the median barely budges. 🍴 Remix it to swap the last tab's four data sets for your own class's numbers.
Sampling Bias
Change who you survey to build your own samples and see why the same question keeps giving a different winner.
Correlation vs Causation
Change temperature and ice-cream sales separately to see that two lines rising together don't prove cause — a hidden factor can move both.
Misleading Graphs
Change the vertical axis start and the time range to see the same data puff up to look 7 times bigger.
Quartiles & Box Plots
Shift and spread the data to see how the minimum, Q1, median, Q3, maximum and the box plot itself change.
Bar and Pie Charts
Show the same data as a strip chart and a pie chart, then change the number of items to see how the percentages shift.
Two-Way Tables
Change the counts of gender × favorite animal in a two-way table to see how marginal totals and relative frequencies are linked.