Pick Your Method

The Method-Selection Map
The Method-Selection Map Start with YOUR question — follow the forks — land on ONE method Data Activism · Data-Analysis Day · The Multiverse School ★ WORKED EXAMPLE — WATCH THE TOOL RUN “Does living closer to the Ironton plant predict whether a Lincoln Elementary kid has asthma — yes / no?” your dataset: EPA AQS air-quality readings + school health records 1 Start with your question 2 Labeled outcome to predict? → YES 3 A number or a category? → CATEGORY 4 Two classes, or many? → TWO (yes / no) 5 Land on it: LOGISTIC REGRESSION STEP 0 · ALWAYS RUN THESE FIRST — before you ask anything below DESCRIBE EDA / descriptive stats What's even in here? Distributions, outliers, gaps. COMPARE Hypothesis testing Is this difference real, or just noise? (t-test / χ²) START WITH YOUR QUESTION What are you trying to learn from YOUR dataset? Write it in one sentence — then follow the forks. 1 ★ OUR WORKED EXAMPLE Do you have a labeled outcome you want to predict? 2 No labels — what are you looking for? Is the outcome a NUMBER or a CATEGORY? 3 NO — unlabeled YES — I have labels CLUSTER k-means clustering Natural groupings we didn't know about REDUCE PCA / dimension reduction See high-dimensional data in 2D hidden groups too many columns REGRESS Linear regression Does X drive Y, and by how much? Two classes, or many? 4 NUMBER CATEGORY CLASSIFY SVM Complex, non-linear boundary between classes CLASSIFY Logistic regression Which factors push a yes / no? 5 RULES Decision trees Sort into several human-readable rules TWO MANY boundary not a straight line? THESE 8 GET YOU STARTED Ask your AI sandbox "what other method fits this question and data shape?" — and see the scikit-learn "choosing the right estimator" map.

Tour mode

Space / → next · ← previous
Home / End first / last
1…9 jump to frame

Explore mode

Click any highlighted region to zoom in.
Click further-in regions to drill deeper.
Backspace / Esc — zoom back out
O — return to overview

Anywhere

M — toggle mode
T — toggle frame list
F — fullscreen · R — restart
0 / 0 Tour