tracki/companies-dataset

Tracki - Synthetic Companies Dataset

10,196 synthetic companies x 18 columns. Every one of the 15 content fields was written by a language model; nothing is rule-generated. Built for the Tracki final project (RUNI - Intro to Data Science): describe a company or an idea, and Tracki returns the 3 most similar companies (embeddings) and generates 1 new competitor profile.

Every row is fictional. Where the generator's name prior collided with a real trademark it is measured and reported below, not denied.

Schema

Cardinalities and ranges are the observed ones, computed from the published file - not the enum the model was offered. Where the two differ, the gap is itself a finding; see Distribution collapse.

column	source	observed in the data	description
name	LLM	10,196 distinct (100.0% of rows)	Fictional company name (1-3 words). Regenerated in a bulk second pass.
tagline	LLM	10,196 distinct (100.0% of rows)	Marketing tagline (2-8 words). Regenerated in a bulk second pass.
value_prop	LLM	10,196 distinct (100.0% of rows)	One-sentence value proposition. The row's identity and the core of the embedding text.
industry	LLM	15 values, top Fintech 8.2%	Sector. STEERED - asked for explicitly, one per batch.
target_customer	LLM	8 values, top Developers 36.8%	Who they sell to.
business_model	LLM	5 values, top B2B 87.3%	B2B / B2C / B2B2C / D2C / B2G.
revenue_model	LLM	5 values, top Subscription 52.8%	How they charge.
deployment_model	LLM	4 values, top Cloud 97.3%	How the product is delivered.
funding_stage	LLM	8 values, top Series A 13.5%	Ordinal, Bootstrapped -> Public. STEERED - asked for explicitly, one per batch.
primary_market	LLM	3 values, top Global 85.7%	Where they sell.
hq_country	LLM	7 values, top USA 99.0%	Headquarters.
founded_year	LLM	2005 - 2024, 20 distinct	Year founded.
num_employees	LLM	4 - 13000, 61 distinct	Headcount. Band-checked against funding_stage.
funding_total_usd_m	LLM	0 - 3000, 76 distinct	Total raised, USD millions. Band-checked against funding_stage.
annual_revenue_usd_m	LLM	0 - 7500, 121 distinct	Annual revenue, USD millions. Band-checked against funding_stage.
company_age	engineered	2 - 21, 20 distinct	2026 - founded_year. Deterministic - see Redundancy.
stage_rank	engineered	0 - 7, 8 distinct	Integer rank of funding_stage. Deterministic - see Redundancy.
embed_text	engineered	10,196 distinct (100.0% of rows)	Legacy concatenation of 7 columns. NOT the text the recommender embeds - Part 3 measured its categorical prefix (39.5% of the string, only 306 distinct values) as harmful and embeds tagline + value_prop instead. See embeddings/eval_report.md.

How it was built

Three generator notebooks ran in parallel Colab tabs (seeds [42, 43, 44]), each owning 5 of the 15 industries, with Qwen/Qwen2.5-3B-Instruct. Prompts were GPU-batched; every batch was validated and filtered in the loop, and any row breaking a bound was dropped, not repaired.

name and tagline were then regenerated by Qwen/Qwen2.5-0.5B-Instruct in a bulk second pass - 50 items per prompt rather than one, so the model sees and avoids its own previous outputs in context. random picks only which industry and which style to request; it never writes a value.

Two of the fifteen content columns were steered - industry and funding_stage were asked for explicitly, one per batch, because an unsteered pilot collapsed onto Series B. The model still had to write the value, and a mismatch was a rejected row. The other thirteen were left to the model's own choice, which is why they look the way the next section describes.

Deduplication

value_prop is the row's identity - it is what gets embedded and what the recommender compares.

step	before	after	removed
exact value_prop	10,810	10,803	-7
normalized value_prop	10,803	10,782	-21
semantic value_prop (>0.94)	10,782	10,196	-586

The semantic pass (cosine > 0.94 over sentence-transformers/all-MiniLM-L6-v2) removes near-twins that exact matching missed - rows differing only by a trailing period, a swapped preposition, or reordered clauses. They matter because each near-twin would be the other's nearest neighbour by construction.

Row accounting

stage	rows
generation attempts	12,436
rejected by the bounds validator	1,626
written to shards	10,810
surviving the dedupe cascade	10,196
published	10,196

• Row-level validity 86.93% - of everything attempted, the fraction satisfying all bounds. A statement about the validator.
• End-to-end yield 81.99% - of everything attempted, the fraction reaching the published file, after deduplication too. A statement about the pipeline.

Note on earlier releases. A previous version published 2,726 rows: the merge step deduplicated on name as well as value_prop, discarding ~8,000 valid rows whose only fault was sharing a company name with an earlier row, which pinned the dataset size to the generator's name vocabulary rather than to how much data was generated. That filter is gone. A later version shipped this file with a card whose enum sizes described the prompt rather than the data; that is what the present card fixes.

Distribution collapse - what the EDA caught

Thirteen of the fifteen content columns were the model's unsteered choice, and a small instruction-tuned model has a narrow prior. It shows:

column	offered	used	levels never used	most common value
hq_country	15	7	8 never used	USA 99.0%
deployment_model	5	4	1 never used	Cloud 97.3%
business_model	5	5	-	B2B 87.3%
primary_market	7	3	4 never used	Global 85.7%
revenue_model	7	5	2 never used	Subscription 52.8%
target_customer	8	8	-	Developers 36.8%
funding_stage	8	8	-	Series A 13.5%
industry	15	15	-	Fintech 8.2%

The two steered columns are flat; every unsteered one is concentrated. hq_country reaches 99.0% on a single value. Some offered levels drew zero rows at all.

What this means for use. Columns above roughly 85% concentration carry almost no information - do not condition on them, stratify by them, or read the region and country charts as geography. The fix for a future run is the one that already worked twice: steer them the way industry and funding_stage were steered, asking for one level per batch and rejecting a mismatch, rather than hoping for diversity from an unconstrained enum.

Grid-lock

column	distinct values	rows per distinct value	most common	top 10 cover
founded_year	20	509.8	2018 21.3%	95.6%
num_employees	61	167.1	15 11.1%	62.4%
funding_total_usd_m	76	134.2	0.5 11.6%	65.6%
annual_revenue_usd_m	121	84.3	5 9.1%	54.2%

The numerics are round numbers, not samples from a distribution. Treat the numeric histograms as bar charts over a small lattice.

Redundancy

company_age is exactly 2026 - founded_year and stage_rank is exactly the rank of funding_stage - Spearman rho between the first pair is -1.000 by construction. They are kept for convenience; do not feed both members of a pair to a model, and do not read their correlation cell as a result. 8.47% of rows are also exact duplicates across all non-text columns.

Consistency - and whether the validator forced it

The model chose every number. Rows outside the stage band were dropped, so the obvious question is whether a rising ladder could have failed to appear. It could:

funding_stage	funding allowed	funding produced	range used	median	distinct	employees produced	n
Bootstrapped	0 - 1	0 - 1	100%	0.5	9	5 - 20	1,182
Pre-seed	0.05 - 2	0.2 - 1.5	67%	1.2	12	4 - 10	1,362
Seed	0.5 - 8	2 - 6	53%	3.5	9	8 - 35	1,358
Series A	5 - 30	10 - 25	60%	15	9	30 - 100	1,378
Series B	20 - 90	40 - 75	50%	50	8	120 - 350	1,377
Series C	60 - 250	100 - 180	42%	150	11	200 - 750	1,366
Series D+	150 - 900	300 - 750	60%	500	11	400 - 2500	1,303
Public	200 - 5000	250 - 3000	57%	1200	19	1000 - 13000	870

24 of 24 adjacent stage pairs have overlapping permitted bands, across all four band-checked numerics - so a non-monotonic median was reachable inside the validator for every one of them. The model did not produce one. It went further and self-separated: 6 of those pairs came out with disjoint realised ranges despite overlapping permission. And it used only 24-100% of the room each band gave it.

Unaided band compliance, measured over all 12,436 attempts:

check	attempts passing
every band at once	95.05%
funding within stage band	97.79%
headcount within stage band	99.58%
revenue within stage band	99.96%
revenue per head plausible	97.74%
age within stage band	99.99%
stage matches its own numbers	99.99%

funding_stage	median funding (M)	median revenue (M)	median employees	n
Bootstrapped	0.50	1.20	12	1,182
Pre-seed	1.20	0.50	7	1,362
Seed	3.50	1.20	15	1,358
Series A	15.00	5.00	50	1,378
Series B	50.00	45.00	150	1,377
Series C	150.00	75.00	350	1,366
Series D+	500.00	420.00	800	1,303
Public	1,200.00	1,500.00	2,350	870

• funding - monotonic Pre-seed->Public: True, Spearman rho = 0.992
• employees - monotonic Pre-seed->Public: True, Spearman rho = 0.99
• revenue - monotonic Pre-seed->Public: True, Spearman rho = 0.982

Bootstrapped is deliberately outside that range: it sits below Pre-seed in stage_rank but legitimately carries more revenue and headcount, so including it would report a failure for a correct model.

The honest counterweight. Each stage holds only 8-19 distinct funding values, so rho = 0.992 partly measures the model snapping to canonical round numbers per stage rather than a genuinely tight relationship. The numeric block behaves as one latent variable: predicting funding_stage from the three numeric columns is near-perfect by construction and is not a meaningful ML task on this data.

Validity - what the model got wrong

Every rejection is a row the model wrote that broke a bound and was dropped, not repaired. All 16 reasons are listed, including the singletons - funding_stage and band_age being near-zero is the point, not a rounding detail.

reason	rows	% of attempts
tagline	331	2.66%
target_customer	301	2.42%
band_rev_per_head	281	2.26%
band_funding	275	2.21%
value_prop	262	2.11%
primary_market	60	0.48%
band_employees	52	0.42%
non_english	22	0.18%
bad_json	14	0.11%
business_model	10	0.08%
revenue_model	9	0.07%
band_revenue	5	0.04%
funding_stage	1	0.01%
deployment_model	1	0.01%
band_age	1	0.01%
hq_country	1	0.01%
total	1,626	13.07%

Known collisions with real company names

The generator's name prior contains real companies. A hand-written blocklist of 168 well-known names finds 10 exact matches (10 rows) and 19 first-word matches (23 rows, 0.23% of the dataset):

Asana, Figma, HubSpot, Instacart, Mixpanel, Mosaic, Salesforce, Slack, Zendesk, Zoom

Those rows are fictional and their figures are fabricated - a row named Slack carries invented headcount and funding and describes an invented product. The blocklist is short and hand-written, so this is a lower bound. Do not treat any name in this dataset as referring to a real organisation. The fix for a future run is to screen generated names against a blocklist at generation time.

What the recommender inherits

• embed_text opens with the sector name in 100.0% of rows, and a row's nearest neighbour shares its industry 98.8% of the time against a 7.1% random baseline. Similarity here is substantially a sector match; compare embedding models on a variant without the boilerplate prefix as well.
• The 0.94 dedupe cut leaves residual near-twins behind:

nearest-neighbour cosine	rows	share
>= 0.95	0	0.00%
>= 0.90	1,336	13.10%
>= 0.85	3,707	36.36%
>= 0.80	5,969	58.54%

Each surviving pair is the other's top-1 recommendation, so a handful of queries will return a near-restatement of the query.

Limitations

• Fully synthetic. Every value was written by a small instruction-tuned LLM and validated only against bounds - plausible-looking numbers are not market data and must not be used for any real decision.
• Thirteen of fifteen content columns were unsteered, so their distributions reflect the model's prior, not the real economy. See Distribution collapse for the measured concentrations.
• Rows violating any bound were dropped rather than repaired, so the kept rows are a biased sample of what the model produces. The reject table describes what was thrown away.
• name and tagline were regenerated in a second bulk pass; in the original single pass the model produced only ~2,726 distinct names across 10,803 rows.
• Real company names appear in name - see the collisions section. Rows are fictional regardless.
• company_age and stage_rank are deterministic functions of other columns.
• Shard 3 delivered 2,385 of its 4,200 target, so Edtech / Foodtech / Mobility / HRtech / Gaming carry fewer rows than the other ten sectors.

Files

• companies_final.parquet / .csv - the dataset (SHA-256 991e84089c0bd426...)

• manifest.json - models, seeds, dedupe cascade, counts, fingerprint

• raw/ - the pre-dedupe shards and per-shard reject logs

• notebooks actually present in this repo:

• notebooks/companies_merge_eda_publish.ipynb

• notebooks/companies_p1.ipynb

• notebooks/companies_p2.ipynb

• notebooks/companies_p3.ipynb

• notebooks/companies_rebuild_v2.ipynb

• notebooks/tracki_embeddings.ipynb

EDA