retentioneering-product-analytics

Name: retentioneering-product-analytics
Author: retentioneering

907stars136forks

How to install retentioneering-product-analytics?

retentioneering/retentioneering-tools/retentioneering-product-analytics

$npx -y skills add retentioneering/retentioneering-tools --skill retentioneering-product-analytics

Installs into the current project.

Summary

Analyze event logs, clickstreams, user paths, product funnels, retention, behavioral segments, transition graphs, step matrices, sequence patterns, and customer journeys using Retentioneering. Use when the user provides CSV, Parquet, pandas, or database event data containing user

Files · 1

View on GitHub

SKILL.md

1	# Retentioneering product analytics
2
3	## Objective
4
5	Turn event-level behavioral data into a reproducible answer to a product question —
6	why users convert, churn, loop, or abandon — using user trajectories, transitions,
7	funnels, and behavioral segments.
8
9	Do not merely generate visualizations. Connect each output to the question, separate
10	observation from interpretation, and never present path correlations as causal effects.
11
12	## Bundled references (read on demand, not upfront)
13
14	\| File \| Read it when \|
15	\|---\|---\|
16	\| `references/api-map.md` \| before writing any Retentioneering call — verified signatures, argument conventions, return shapes for 5.x \|
17	\| `references/analysis-recipes.md` \| after the question is clear — 10 field-tested patterns (R1–R10) with skeletons and pitfalls \|
18	\| `references/gotchas-and-validation.md` \| before executing (API gotchas G1–G10) and before presenting (integrity checklist B1–B10) \|
19	\| `scripts/inspect_event_log.py` \| step 2 — automated data profiling and schema suggestion \|
20
21	## Required event-log semantics
22
23	Minimum: a path identifier (user or session), an event name, a timestamp (or a reliable
24	order column — see gotcha G2 for order-only data). Useful extras: session id, segment
25	attributes (device, source, plan), event properties, conversion labels.
26
27	## Workflow
28
29	### 1. Environment
30
31	1. Confirm the package: `python -c "import retentioneering; print(retentioneering.__version__)"`.
32	Expect 5.x; this skill's API map is version-verified for 5.0 — on a different major
33	version, trust installed docstrings over the map.
34	2. Locate the event data (CSV / Parquet / frames in existing code). Never modify inputs.
35	3. Do not invent methods: anything not in `references/api-map.md` must be verified
36	against the installed package before use.
37
38	### 2. Inspect the data BEFORE choosing methods
39
40	Run `scripts/inspect_event_log.py <path> [--sep ...]` (or replicate its checks inline for
41	in-memory frames). It profiles columns, infers the user/event/timestamp mapping, checks
42	timestamp parseability, duplicates, per-path ordering, path-length distribution, and
43	emits `artifacts/data-profile.json` plus a ready-to-paste `Eventstream(...)` schema.
44
45	Report to the user before proceeding: inferred mapping, row/user/event-type counts,
46	covered period, and any red flags (nulls in key columns, timestamp ties, suspected bots
47	or ultra-long paths, order-only timestamps). Confirm the mapping if inference is
48	ambiguous.
49
50	### 3. Frame the product question, then pick the SMALLEST recipe
51
52	Map the question to a recipe in `references/analysis-recipes.md`:
53	navigation structure/loops → transition graph (R1/R6) · before/after an anchor →
54	step matrix (R3) · ordered conversion flow → funnel (R1/R2) · what winners do
55	differently → diff on a funnel-stage segment (R2) · heterogeneous users → clustering
56	without target leakage (R5) · between two funnel levels → truncate micro-journey (R4) ·
57	intervention timing → time-to-outcome (R7) · cross-segment scan → segment overview (R8) ·
58	value of a fix → Markov what-if (R9, advanced).
59
60	Combine recipes only when each addition resolves a distinct uncertainty.
61
62	### 4. Execute reproducibly
63
64	1. Prefer a rerunnable script (or a notebook executed top-to-bottom) over ad-hoc cells.
65	2. Write artifacts to a dedicated output directory (`artifacts/` by default).
66	3. Log every filtering rule and its row/path impact; never silently drop data
67	(integrity item B2).
68	4. Use `sample_paths(frac=, random_state=)` for stable subsamples; stochastic steps get
69	explicit seeds.
70	5. Record lineage: save `processed.recipe()` and the package version into
71	`artifacts/run-metadata.json` — any artifact must be regenerable from raw data vi

People who liked this also viewed

6 picks

Type

retentioneering-product-analytics

907stars136forks

View on GitHub

How to install retentioneering-product-analytics?

retentioneering/retentioneering-tools/retentioneering-product-analytics

$npx -y skills add retentioneering/retentioneering-tools --skill retentioneering-product-analytics

Installs into the current project.

Summary

Files · 1

View on GitHub

SKILL.md

1	# Retentioneering product analytics
2
3	## Objective
4
5	Turn event-level behavioral data into a reproducible answer to a product question —
6	why users convert, churn, loop, or abandon — using user trajectories, transitions,
7	funnels, and behavioral segments.
8
9	Do not merely generate visualizations. Connect each output to the question, separate
10	observation from interpretation, and never present path correlations as causal effects.
11
12	## Bundled references (read on demand, not upfront)
13
14	\| File \| Read it when \|
15	\|---\|---\|
16	\| `references/api-map.md` \| before writing any Retentioneering call — verified signatures, argument conventions, return shapes for 5.x \|
17	\| `references/analysis-recipes.md` \| after the question is clear — 10 field-tested patterns (R1–R10) with skeletons and pitfalls \|
18	\| `references/gotchas-and-validation.md` \| before executing (API gotchas G1–G10) and before presenting (integrity checklist B1–B10) \|
19	\| `scripts/inspect_event_log.py` \| step 2 — automated data profiling and schema suggestion \|
20
21	## Required event-log semantics
22
23	Minimum: a path identifier (user or session), an event name, a timestamp (or a reliable
24	order column — see gotcha G2 for order-only data). Useful extras: session id, segment
25	attributes (device, source, plan), event properties, conversion labels.
26
27	## Workflow
28
29	### 1. Environment
30
31	1. Confirm the package: `python -c "import retentioneering; print(retentioneering.__version__)"`.
32	Expect 5.x; this skill's API map is version-verified for 5.0 — on a different major
33	version, trust installed docstrings over the map.
34	2. Locate the event data (CSV / Parquet / frames in existing code). Never modify inputs.
35	3. Do not invent methods: anything not in `references/api-map.md` must be verified
36	against the installed package before use.
37
38	### 2. Inspect the data BEFORE choosing methods
39
40	Run `scripts/inspect_event_log.py <path> [--sep ...]` (or replicate its checks inline for
41	in-memory frames). It profiles columns, infers the user/event/timestamp mapping, checks
42	timestamp parseability, duplicates, per-path ordering, path-length distribution, and
43	emits `artifacts/data-profile.json` plus a ready-to-paste `Eventstream(...)` schema.
44
45	Report to the user before proceeding: inferred mapping, row/user/event-type counts,
46	covered period, and any red flags (nulls in key columns, timestamp ties, suspected bots
47	or ultra-long paths, order-only timestamps). Confirm the mapping if inference is
48	ambiguous.
49
50	### 3. Frame the product question, then pick the SMALLEST recipe
51
52	Map the question to a recipe in `references/analysis-recipes.md`:
53	navigation structure/loops → transition graph (R1/R6) · before/after an anchor →
54	step matrix (R3) · ordered conversion flow → funnel (R1/R2) · what winners do
55	differently → diff on a funnel-stage segment (R2) · heterogeneous users → clustering
56	without target leakage (R5) · between two funnel levels → truncate micro-journey (R4) ·
57	intervention timing → time-to-outcome (R7) · cross-segment scan → segment overview (R8) ·
58	value of a fix → Markov what-if (R9, advanced).
59
60	Combine recipes only when each addition resolves a distinct uncertainty.
61
62	### 4. Execute reproducibly
63
64	1. Prefer a rerunnable script (or a notebook executed top-to-bottom) over ad-hoc cells.
65	2. Write artifacts to a dedicated output directory (`artifacts/` by default).
66	3. Log every filtering rule and its row/path impact; never silently drop data
67	(integrity item B2).
68	4. Use `sample_paths(frac=, random_state=)` for stable subsamples; stochastic steps get
69	explicit seeds.
70	5. Record lineage: save `processed.recipe()` and the package version into
71	`artifacts/run-metadata.json` — any artifact must be regenerable from raw data vi

People who liked this also viewed

6 picks

Type