User 136 · Retail · Baskets · Association Rules
Executive Summary

Executive Summary

Key findings from the market basket analysis

Transactions
820
Unique Items
1124
Rules Discovered
8165
Avg Lift
444.037
Max Lift
820
Min Support
0.001
Min Confidence
0.3
Across 820 transactions and 1124 unique products, the Apriori algorithm discovered 8165 association rules. The strongest association has a lift of 820, meaning those products are bought together 820x more often than chance would predict. The average lift across all rules is 444.04, indicating consistent cross-selling opportunities throughout the catalogue.
Interpretation

Across 820 transactions and 1124 unique products, the Apriori algorithm discovered 8165 association rules. The strongest association has a lift of 820, meaning those products are bought together 820x more often than chance would predict. The average lift across all rules is 444.04, indicating consistent cross-selling opportunities throughout the catalogue.

Visualization

Most Frequently Purchased Items

Top 30 products ranked by transaction support

Interpretation

The most frequently purchased item is 'JUMBO BAG RED RETROSPOT', appearing in 1.5% of all transactions. These high-frequency items act as anchor products: they generate the most rules and appear on both sides of many associations. High item support alone doesn't mean a product has strong affinities — lift accounts for this popularity bias when evaluating rule strength.

Data Table

Top Association Rules by Lift

Top 20 rules sorted by lift (strongest associations first)

LiftCountSupportConfidenceAntecedentsConsequents
82010.00121PINK DINER WALL CLOCKEDWARDIAN PARASOL RED
82010.00121EDWARDIAN PARASOL REDPINK DINER WALL CLOCK
82010.00121MAGIC DRAWING SLATE LEAP FROGMINI JIGSAW DOLLY GIRL
82010.00121MINI JIGSAW DOLLY GIRLMAGIC DRAWING SLATE LEAP FROG
82010.0012136 PENCILS TUBE RED RETROSPOTCHILDREN'S CIRCUS PARADE MUG
82010.00121CHILDREN'S CIRCUS PARADE MUG36 PENCILS TUBE RED RETROSPOT
82010.00121PACK OF 60 DINOSAUR CAKE CASESWHITE WICKER STAR
82010.00121WHITE WICKER STARPACK OF 60 DINOSAUR CAKE CASES
82010.00121HANGING BAUBLE T-LIGHT HOLDER SMALLPENS ASSORTED SPACEBALL
82010.00121PENS ASSORTED SPACEBALLHANGING BAUBLE T-LIGHT HOLDER SMALL
82010.00121CERAMIC CHERRY CAKE MONEY BANKEDWARDIAN PARASOL BLACK
82010.00121EDWARDIAN PARASOL BLACKCERAMIC CHERRY CAKE MONEY BANK
82010.00121CHRISTMAS RETROSPOT ANGEL WOODRECIPE BOX RETROSPOT
82010.00121RECIPE BOX RETROSPOTCHRISTMAS RETROSPOT ANGEL WOOD
82010.00121MADRAS NOTEBOOK MEDIUMVANILLA SCENT CANDLE JEWELLED BOX
82010.00121VANILLA SCENT CANDLE JEWELLED BOXMADRAS NOTEBOOK MEDIUM
82010.00121RUSTIC STRAWBERRY JAM POT LARGESET 12 COLOURING PENCILS DOILY
82010.00121SET 12 COLOURING PENCILS DOILYRUSTIC STRAWBERRY JAM POT LARGE
82010.00121SCENTED VELVET LOUNGE CANDLESET OF 6 FUNKY BEAKERS
82010.00121SET OF 6 FUNKY BEAKERSSCENTED VELVET LOUNGE CANDLE
Interpretation

The highest-lift rule links 'PINK DINER WALL CLOCK' to 'EDWARDIAN PARASOL RED' with a lift of 820 and 100% confidence. Lift above 1 confirms the association is non-random — these products are genuinely more likely to be purchased together than their individual popularity would suggest. Rules are sorted by lift so the most actionable cross-sell pairs appear first.

Visualization

Support vs. Confidence of All Rules

Scatter of all discovered rules; bubble size = lift

Interpretation

Each point represents one association rule plotted by its support (x-axis) and confidence (y-axis), with bubble size proportional to lift. 8165 rules are shown in total; the 2324 highest-lift rules (top 10% by lift, lift >= 820) are the most actionable. Rules in the upper-left region are reliable but niche (high confidence, low support), while upper-right rules are both common and reliable cross-sell opportunities.

Visualization

Distribution of Rule Lift Values

Histogram of lift values across all discovered rules

Interpretation

The lift distribution shows that most rules cluster around an average of 444.04, with the strongest rule reaching 820. 90% of rules have a lift below 820, meaning a small number of high-lift rules account for the most powerful associations. A right-skewed distribution is typical: many rules with modest lift and a long tail of exceptional product affinities that deserve priority attention in merchandising.

Visualization

Top Recommended Products (Consequents by Avg Lift)

Products most frequently recommended with highest average lift

Interpretation

Products that appear as rule consequents with high average lift are prime cross-sell targets. '12 PENCIL SMALL TUBE WOODLAND' leads with an average lift of 820 across all rules where it appears as the predicted purchase. These products should be prominently featured in recommendation widgets, product-detail page suggestions, and checkout prompts alongside their associated antecedents.

Data Table

Rule Set Statistics

Aggregate metrics describing the full discovered rule set

MetricValue
Total Rules Discovered8165
Transactions Analysed820
Unique Products1124
Min Support Threshold0.001
Min Confidence Threshold0.3
Average Lift444.037
Maximum Lift820
Interpretation

The analysis examined 820 multi-item transactions and produced 8165 association rules meeting the configured thresholds. These statistics characterise the overall richness of the rule set: a high average lift indicates that most discovered rules reflect genuine product affinities rather than mere coincidence driven by item popularity. Adjusting min_support and min_confidence tunes the breadth-depth tradeoff — lower thresholds yield more rules but may include weaker associations.

Your data has more stories to tell. Run any analysis on your own data — 60+ validated R modules, interactive reports, AI insights, and PDF export. 2,000 free credits on signup.
Try Free — No Signup Sign Up Free

Report an Issue

Tell us what's wrong. You'll get a free re-run of this analysis so you can try again with different parameters. If the re-run still doesn't meet your expectations, we'll refund your credits.

Want to run this analysis on your own data? Upload CSV — Free Analysis See Pricing