{
"cells": [
{
"cell_type": "markdown",
"id": "58e6f57f",
"metadata": {},
"source": [
"# Custom Dataset From Scratch\n",
"\n",
"Most of the other tutorials start from a **reference** risk dataset that\n",
"Bayesline maintains — `bayesline/Bayesline-US-All-1y`, `Bayesline-Global`,\n",
"etc. You layer your own exposures or filters on top via\n",
"`DerivedRiskDatasetSettings` and the engine pulls everything else (master\n",
"data, prices, calendars, FX) from the reference.\n",
"\n",
"This tutorial covers the other path: **building a risk dataset from scratch\n",
"with no reference**. You bring every piece of input yourself. This is the\n",
"path you take when you have your own house-alpha factors, your own price\n",
"feed, your own asset master, and you want Bayesline to fit a factor model\n",
"on top of all of it without any external data dependency.\n",
"\n",
"The mechanism is `RootRiskDatasetSettings` and the six (+1 optional) upload\n",
"data types that feed it:\n",
"\n",
"```{mermaid}\n",
"erDiagram\n",
" idmap ||--o{ market_cap : \"asset_id\"\n",
" idmap ||--o{ price : \"asset_id\"\n",
" idmap ||--o{ exposures : \"asset_id\"\n",
" exchange_rates ||--o{ market_cap : \"ccy\"\n",
" exchange_rates ||--o{ price : \"ccy\"\n",
"\n",
" idmap {\n",
" date start_date\n",
" date end_date\n",
" string from_id_type\n",
" string from_id\n",
" string to_id_type\n",
" string to_id\n",
" }\n",
" market_cap {\n",
" date date PK\n",
" string asset_id PK\n",
" string asset_id_type PK\n",
" string ccy\n",
" float market_cap\n",
" float volume\n",
" float idio_vol\n",
" }\n",
" price {\n",
" date date PK\n",
" string asset_id PK\n",
" string asset_id_type PK\n",
" string ccy PK\n",
" float close\n",
" float return\n",
" bool delisted\n",
" }\n",
" exposures {\n",
" date date PK\n",
" string asset_id PK\n",
" string asset_id_type PK\n",
" string factor_group PK\n",
" string factor PK\n",
" float exposure\n",
" }\n",
" exchange_dates {\n",
" date date PK\n",
" string exchange PK\n",
" }\n",
" exchange_rates {\n",
" date date PK\n",
" string ccy PK\n",
" float fx_rate\n",
" }\n",
"```\n",
"\n",
"One sentence per box:\n",
"\n",
"* `idmap` — *which ids refer to the same asset?* (ticker → ticker_core,\n",
" ISIN → ticker, etc.)\n",
"* `market_cap` — *who is each asset?* Slow-changing master data: id, ccy,\n",
" market cap, volume, idio vol.\n",
"* `price` — *what did each asset do today?* The daily market record:\n",
" close, daily return, delisted flag.\n",
"* `exposures` — *what factors does each asset load on?* Long format:\n",
" `(date, asset, group, factor, value)`.\n",
"* `exchange_dates` — *which days is each exchange closed?* Non-trading\n",
" days, per exchange.\n",
"* `exchange_rates` — *how do we get to USD?* USD-base FX per `(date, ccy)`.\n",
"\n",
"The rest of this tutorial walks one box at a time, simulates a small US\n",
"universe so everything runs offline, builds a dataset from all six\n",
"uploads, fits a factor model on top, and checks that the engine recovers\n",
"the factor returns we simulated.\n",
"\n",
"If you already have a reference dataset that meets your needs, you want\n",
"[Model Onboarding](recipe_model_onboarding.ipynb) instead — same final\n",
"shape, derived path.\n"
]
},
{
"cell_type": "markdown",
"id": "58e22559",
"metadata": {},
"source": [
"## Imports and client\n",
"\n",
"We pull in `polars` for frame construction, `numpy` for the simulation,\n",
"and the public `bayesline.api.equity` settings types we'll need below.\n"
]
},
{
"cell_type": "code",
"execution_count": 2,
"id": "87d632b5",
"metadata": {
"lines_to_next_cell": 2
},
"outputs": [],
"source": [
"import datetime as dt\n",
"\n",
"import numpy as np\n",
"import polars as pl\n",
"\n",
"from bayesline.apiclient import BayeslineApiClient\n",
"from bayesline.api.equity import (\n",
" CategoricalExposureGroupSettings,\n",
" CategoricalFilterSettings,\n",
" ContinuousExposureGroupSettings,\n",
" ExposureSettings,\n",
" FactorRiskModelSettings,\n",
" ModelConstructionSettings,\n",
" RootRiskDatasetSettings,\n",
" UniverseSettings,\n",
")"
]
},
{
"cell_type": "markdown",
"id": "589a150b",
"metadata": {},
"source": [
"A real script would connect to your deployment via\n",
"`BayeslineApiClient.new_client(endpoint=..., api_key=...)`. This tutorial\n",
"uses the in-process app the docs build provides — every API call below\n",
"runs end-to-end against the real backend.\n"
]
},
{
"cell_type": "code",
"execution_count": null,
"id": "88adc30a",
"metadata": {
"lines_to_next_cell": 2,
"tags": [
"skip-execution"
]
},
"outputs": [],
"source": [
"bln = BayeslineApiClient.new_client(\n",
" endpoint=\"https://[ENDPOINT]\",\n",
" api_key=\"[API-KEY]\",\n",
")"
]
},
{
"cell_type": "markdown",
"id": "a1db937f",
"metadata": {},
"source": [
"## Simulating a universe\n",
"\n",
"To keep the tutorial self-contained we generate everything in-notebook:\n",
"30 assets on the NYSE, one year of business days, six industries, three\n",
"style factors, plus a market intercept. Every input frame downstream is\n",
"derived from the constants in this cell.\n",
"\n",
"The simulation also gives us **ground truth**: we draw factor returns\n",
"ourselves, so at the end we can plot the engine's estimated `fret()`\n",
"against what we know the answer should be.\n"
]
},
{
"cell_type": "code",
"execution_count": 3,
"id": "bc1add0b",
"metadata": {
"lines_to_next_cell": 2
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"30 assets × 261 dates (2024-01-02 → 2024-12-31)\n",
"industry counts: {np.str_('CONSUMER'): np.int64(8), np.str_('ENERGY'): np.int64(2), np.str_('FINS'): np.int64(6), np.str_('HEALTH'): np.int64(6), np.str_('MATERIALS'): np.int64(4), np.str_('TECH'): np.int64(4)}\n"
]
}
],
"source": [
"rng = np.random.default_rng(42)\n",
"\n",
"N_ASSETS = 30\n",
"INDUSTRIES = [\"TECH\", \"ENERGY\", \"FINS\", \"HEALTH\", \"CONSUMER\", \"MATERIALS\"]\n",
"STYLES = [\"momentum\", \"value\", \"size\"]\n",
"\n",
"# 1 year of weekdays ending 2024-12-31. We filter weekends here so the\n",
"# simulation doesn't generate prices on Sat/Sun; both weekends and US\n",
"# holidays are declared non-trading in the `exchange_dates` upload\n",
"# below (the engine's calendar contract).\n",
"all_days = pl.date_range(\n",
" dt.date(2024, 1, 2), dt.date(2024, 12, 31), interval=\"1d\", eager=True\n",
").to_list()\n",
"dates: list[dt.date] = [d for d in all_days if d.weekday() < 5]\n",
"T = len(dates)\n",
"\n",
"# Stable, readable asset ids. In production these would be tickers, ISINs,\n",
"# or your house ids — anything as long as it's a stable string.\n",
"assets = [f\"A{i:04d}\" for i in range(N_ASSETS)]\n",
"\n",
"# Per-asset static profile: which industry, what style loadings, what base\n",
"# market cap. These are the \"who is each asset\" attributes that flow into\n",
"# market_cap and exposures.\n",
"asset_industry = rng.choice(INDUSTRIES, size=N_ASSETS)\n",
"asset_style_loadings = rng.standard_normal((N_ASSETS, len(STYLES)))\n",
"asset_base_mcap = rng.lognormal(mean=22.0, sigma=1.0, size=N_ASSETS).astype(\"float32\")\n",
"\n",
"print(f\"{N_ASSETS} assets × {T} dates ({dates[0]} → {dates[-1]})\")\n",
"print(\"industry counts:\", dict(zip(*np.unique(asset_industry, return_counts=True))))"
]
},
{
"cell_type": "markdown",
"id": "f58aeac8",
"metadata": {},
"source": [
"### Ground-truth factor returns\n",
"\n",
"We draw factor returns directly, then build per-asset returns as\n",
"\n",
"$$\n",
"r_{i,t} \\;=\\; f^{\\text{market}}_t\n",
" \\;+\\; f^{\\text{industry}}_{g(i),\\,t}\n",
" \\;+\\; \\sum_k \\beta^{\\text{style}}_{i,k}\\, f^{\\text{style}}_{k,t}\n",
" \\;+\\; \\varepsilon_{i,t}.\n",
"$$\n",
"\n",
"Industry returns are constrained to a mcap-weighted zero-sum across\n",
"industries each day — the same constraint we'll apply in the regression\n",
"below. That keeps the industry block orthogonal to the market intercept,\n",
"which is what the engine assumes when `zero_sum_constraints={\"industry\":\n",
"\"mcap_weighted\"}` is set.\n"
]
},
{
"cell_type": "code",
"execution_count": 4,
"id": "24258c7b",
"metadata": {
"lines_to_next_cell": 2
},
"outputs": [
{
"data": {
"text/html": [
"
\n",
"shape: (5, 3)
date
factor
true_return
date
str
f32
2024-01-02
"market.market"
-0.009656
2024-01-03
"market.market"
0.003635
2024-01-04
"market.market"
0.008881
2024-01-05
"market.market"
0.020467
2024-01-08
"market.market"
0.029639
"
],
"text/plain": [
"shape: (5, 3)\n",
"┌────────────┬───────────────┬─────────────┐\n",
"│ date ┆ factor ┆ true_return │\n",
"│ --- ┆ --- ┆ --- │\n",
"│ date ┆ str ┆ f32 │\n",
"╞════════════╪═══════════════╪═════════════╡\n",
"│ 2024-01-02 ┆ market.market ┆ -0.009656 │\n",
"│ 2024-01-03 ┆ market.market ┆ 0.003635 │\n",
"│ 2024-01-04 ┆ market.market ┆ 0.008881 │\n",
"│ 2024-01-05 ┆ market.market ┆ 0.020467 │\n",
"│ 2024-01-08 ┆ market.market ┆ 0.029639 │\n",
"└────────────┴───────────────┴─────────────┘"
]
},
"execution_count": 4,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"SIGMA_MARKET = 0.010 # ~16% annualized\n",
"SIGMA_INDUSTRY = 0.006\n",
"SIGMA_STYLE = 0.004\n",
"SIGMA_IDIO = 0.015 # ~24% annualized\n",
"\n",
"# Market intercept: one return per day.\n",
"f_market = rng.normal(0.0005, SIGMA_MARKET, size=T).astype(\"float32\")\n",
"\n",
"# Industry returns: shape (T, n_industries), then mcap-weighted-zero-summed.\n",
"ind_mcap = np.zeros(len(INDUSTRIES), dtype=\"float32\")\n",
"for k, g in enumerate(INDUSTRIES):\n",
" ind_mcap[k] = asset_base_mcap[asset_industry == g].sum()\n",
"ind_weights = ind_mcap / ind_mcap.sum()\n",
"\n",
"f_industry_raw = rng.normal(0.0, SIGMA_INDUSTRY, size=(T, len(INDUSTRIES))).astype(\n",
" \"float32\"\n",
")\n",
"f_industry = f_industry_raw - (f_industry_raw @ ind_weights)[:, None]\n",
"# Sanity: each day's mcap-weighted industry return is ~0.\n",
"assert np.allclose((f_industry @ ind_weights), 0.0, atol=1e-6)\n",
"\n",
"# Style returns: shape (T, n_styles).\n",
"f_style = rng.normal(0.0, SIGMA_STYLE, size=(T, len(STYLES))).astype(\"float32\")\n",
"\n",
"# Per-asset realized returns: shape (T, N_ASSETS).\n",
"ind_idx = np.array([INDUSTRIES.index(g) for g in asset_industry])\n",
"r = (\n",
" f_market[:, None]\n",
" + f_industry[:, ind_idx]\n",
" + f_style @ asset_style_loadings.T\n",
" + rng.normal(0.0, SIGMA_IDIO, size=(T, N_ASSETS)).astype(\"float32\")\n",
").astype(\"float32\")\n",
"\n",
"# Stash truth as a tidy frame so the final tie-out chart can join against\n",
"# fret(). fret labels factors as `{hierarchy}.{factor}` — `market.market`,\n",
"# `industry.TECH`, `style.momentum`, etc. — so we follow the same scheme\n",
"# here.\n",
"df_truth = pl.concat(\n",
" [\n",
" pl.DataFrame(\n",
" {\"date\": dates, \"factor\": \"market.market\", \"true_return\": f_market}\n",
" ),\n",
" *[\n",
" pl.DataFrame(\n",
" {\n",
" \"date\": dates,\n",
" \"factor\": f\"industry.{g}\",\n",
" \"true_return\": f_industry[:, k],\n",
" }\n",
" )\n",
" for k, g in enumerate(INDUSTRIES)\n",
" ],\n",
" *[\n",
" pl.DataFrame(\n",
" {\"date\": dates, \"factor\": f\"style.{s}\", \"true_return\": f_style[:, k]}\n",
" )\n",
" for k, s in enumerate(STYLES)\n",
" ],\n",
" ]\n",
")\n",
"df_truth.head()"
]
},
{
"cell_type": "markdown",
"id": "be29a9a5",
"metadata": {},
"source": [
"## Step 1 — `idmap`\n",
"\n",
"`idmap` is a temporal table of identifier mappings. It does two jobs in\n",
"a root dataset:\n",
"\n",
"1. **Resolve foreign ids** — ISIN, SEDOL, CUSIP, your house id —\n",
" onto your canonical **master id** at query time.\n",
"2. **Collapse share classes and cross-listings** onto a single\n",
" **master_core id** so the engine treats them as the same underlying\n",
" entity for exposures, market-cap aggregation, and portfolio\n",
" netting.\n",
"\n",
"Job (2) is what the **master vs master_core** distinction is for.\n",
"`master` is your day-to-day asset id — one row per *listing*.\n",
"`master_core` is the canonical id per *entity* — one row per\n",
"underlying company. They are the same id type when every listing is\n",
"its own entity, and they diverge when one entity has multiple\n",
"listings.\n",
"\n",
"The textbook example is Alphabet, which trades as two NASDAQ tickers:\n",
"\n",
"* `GOOGL` — Class A, voting shares\n",
"* `GOOG` — Class C, non-voting shares\n",
"\n",
"Two listings, one company. To bring both into the dataset and collapse\n",
"them onto the same entity, you pick one as the core (say `GOOG`) and\n",
"upload **one projection row per master id**:\n",
"\n",
"| from_id_type | from_id | to_id_type | to_id |\n",
"| ------------ | ------- | ------------- | ------ |\n",
"| `ticker` | `GOOGL` | `ticker_core` | `GOOG` |\n",
"| `ticker` | `GOOG` | `ticker_core` | `GOOG` |\n",
"\n",
"The second row is the **identity projection** for `GOOG` — it tells\n",
"the engine that `GOOG` itself participates as a master id in your\n",
"dataset (and is its own core). Without it, `GOOG` would only be known\n",
"as the *target* of `GOOGL`'s projection, and any `market_cap` /\n",
"`price` / `exposures` row keyed by `GOOG` would be silently dropped\n",
"during dataset construction.\n",
"\n",
"After both rows are in place, `GOOGL` and `GOOG` share one set of\n",
"factor exposures and net to one position when a portfolio is\n",
"aggregated up — even though they continue to carry independent\n",
"prices and market caps in the `price` and `market_cap` uploads.\n",
"\n",
"**A quick warning about using `ticker` as a master_core type.** We\n",
"use `ticker_core` in this tutorial because tickers read well, but\n",
"it's a poor choice in production: tickers can change for the same\n",
"company over time. Facebook re-ticker'd from `FB` to `META` in 2022,\n",
"and any dataset using `ticker_core` as its master_core would split\n",
"that history into two entities — breaking time-series exposures,\n",
"returns, and any portfolio that held it across the rename. A good\n",
"`master_core` id is **permanent** across renames, splits, and\n",
"class consolidations: a vendor permanent id (PERMID, FIGI compid,\n",
"FactSet entity id) or your house's stable id. Bayesline's own\n",
"datasets use `bayesid_core`.\n",
"\n",
"Our toy universe has no share-class or cross-listing pairs, so each\n",
"asset is its own core (the projection is identity). The validity\n",
"window `[0001-01-01, 9999-12-31]` keeps each row valid under any\n",
"`as_of_date`.\n"
]
},
{
"cell_type": "code",
"execution_count": 5,
"id": "e0d970c0",
"metadata": {
"lines_to_next_cell": 2
},
"outputs": [
{
"data": {
"text/html": [
"
\n",
"shape: (5, 6)
start_date
end_date
from_id_type
to_id_type
from_id
to_id
date
date
str
str
str
str
0001-01-01
9999-12-31
"ticker"
"ticker_core"
"A0000"
"A0000"
0001-01-01
9999-12-31
"ticker"
"ticker_core"
"A0001"
"A0001"
0001-01-01
9999-12-31
"ticker"
"ticker_core"
"A0002"
"A0002"
0001-01-01
9999-12-31
"ticker"
"ticker_core"
"A0003"
"A0003"
0001-01-01
9999-12-31
"ticker"
"ticker_core"
"A0004"
"A0004"
"
],
"text/plain": [
"shape: (5, 6)\n",
"┌────────────┬────────────┬──────────────┬─────────────┬─────────┬───────┐\n",
"│ start_date ┆ end_date ┆ from_id_type ┆ to_id_type ┆ from_id ┆ to_id │\n",
"│ --- ┆ --- ┆ --- ┆ --- ┆ --- ┆ --- │\n",
"│ date ┆ date ┆ str ┆ str ┆ str ┆ str │\n",
"╞════════════╪════════════╪══════════════╪═════════════╪═════════╪═══════╡\n",
"│ 0001-01-01 ┆ 9999-12-31 ┆ ticker ┆ ticker_core ┆ A0000 ┆ A0000 │\n",
"│ 0001-01-01 ┆ 9999-12-31 ┆ ticker ┆ ticker_core ┆ A0001 ┆ A0001 │\n",
"│ 0001-01-01 ┆ 9999-12-31 ┆ ticker ┆ ticker_core ┆ A0002 ┆ A0002 │\n",
"│ 0001-01-01 ┆ 9999-12-31 ┆ ticker ┆ ticker_core ┆ A0003 ┆ A0003 │\n",
"│ 0001-01-01 ┆ 9999-12-31 ┆ ticker ┆ ticker_core ┆ A0004 ┆ A0004 │\n",
"└────────────┴────────────┴──────────────┴─────────────┴─────────┴───────┘"
]
},
"execution_count": 5,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"df_idmap = pl.DataFrame(\n",
" {\n",
" \"start_date\": [dt.date.min] * N_ASSETS,\n",
" \"end_date\": [dt.date.max] * N_ASSETS,\n",
" \"from_id_type\": [\"ticker\"] * N_ASSETS,\n",
" \"to_id_type\": [\"ticker_core\"] * N_ASSETS,\n",
" \"from_id\": assets,\n",
" \"to_id\": assets,\n",
" }\n",
")\n",
"df_idmap.head()"
]
},
{
"cell_type": "markdown",
"id": "bc8f39bc",
"metadata": {},
"source": [
"## Step 2 — `market_cap` (entity-level data)\n",
"\n",
"`market_cap` answers *who is each asset?* It carries the asset's id,\n",
"functional currency, market cap, daily volume, and an idiosyncratic-vol\n",
"estimate. Primary key is `(date, asset_id, asset_id_type)`.\n",
"\n",
"**Why is this separate from `price`?** Different levels of aggregation.\n",
"`market_cap` (and the rest of this row — `volume`, `idio_vol`) is an\n",
"**entity-level** quantity: one number per company per day. `price` is\n",
"a **listing-level** quantity: one number per traded ticker per\n",
"currency. Alphabet has two listings — `GOOGL` and `GOOG` — that\n",
"quote at different prices but share a single Alphabet market cap.\n",
"The convention is to upload one `market_cap` row per entity, on the\n",
"master listing (here `GOOG`, whose master id equals its master_core);\n",
"repeating it on every listing would double-count when the engine\n",
"aggregates over the master_core collapse.\n",
"\n",
"In the simulation we build a static per-asset profile, then broadcast\n",
"across dates with a small random-walk perturbation on market cap so the\n",
"frame isn't perfectly constant. `volume` and `idio_vol` are constant in\n",
"this toy — production pipelines fill these from a vendor feed.\n"
]
},
{
"cell_type": "code",
"execution_count": 6,
"id": "dc45e840",
"metadata": {
"lines_to_next_cell": 2
},
"outputs": [
{
"data": {
"text/html": [
"
\n",
"shape: (5, 7)
date
asset_id
asset_id_type
ccy
market_cap
volume
idio_vol
date
str
str
str
f32
f32
f32
2024-01-02
"A0000"
"ticker"
"USD"
2.9318e9
2.655621e7
0.015
2024-01-02
"A0001"
"ticker"
"USD"
1.3932e9
3.0754266e7
0.015
2024-01-02
"A0002"
"ticker"
"USD"
2.5720e9
2.1954332e7
0.015
2024-01-02
"A0003"
"ticker"
"USD"
8.3242e9
3.8139168e7
0.015
2024-01-02
"A0004"
"ticker"
"USD"
6.36904e8
2.5839e6
0.015
"
],
"text/plain": [
"shape: (5, 7)\n",
"┌────────────┬──────────┬───────────────┬─────┬────────────┬─────────────┬──────────┐\n",
"│ date ┆ asset_id ┆ asset_id_type ┆ ccy ┆ market_cap ┆ volume ┆ idio_vol │\n",
"│ --- ┆ --- ┆ --- ┆ --- ┆ --- ┆ --- ┆ --- │\n",
"│ date ┆ str ┆ str ┆ str ┆ f32 ┆ f32 ┆ f32 │\n",
"╞════════════╪══════════╪═══════════════╪═════╪════════════╪═════════════╪══════════╡\n",
"│ 2024-01-02 ┆ A0000 ┆ ticker ┆ USD ┆ 2.9318e9 ┆ 2.655621e7 ┆ 0.015 │\n",
"│ 2024-01-02 ┆ A0001 ┆ ticker ┆ USD ┆ 1.3932e9 ┆ 3.0754266e7 ┆ 0.015 │\n",
"│ 2024-01-02 ┆ A0002 ┆ ticker ┆ USD ┆ 2.5720e9 ┆ 2.1954332e7 ┆ 0.015 │\n",
"│ 2024-01-02 ┆ A0003 ┆ ticker ┆ USD ┆ 8.3242e9 ┆ 3.8139168e7 ┆ 0.015 │\n",
"│ 2024-01-02 ┆ A0004 ┆ ticker ┆ USD ┆ 6.36904e8 ┆ 2.5839e6 ┆ 0.015 │\n",
"└────────────┴──────────┴───────────────┴─────┴────────────┴─────────────┴──────────┘"
]
},
"execution_count": 6,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"# Per-asset static profile (one row per asset).\n",
"df_assets = pl.DataFrame(\n",
" {\n",
" \"asset_id\": assets,\n",
" \"base_mcap\": asset_base_mcap,\n",
" \"volume\": rng.uniform(5e5, 5e7, N_ASSETS).astype(\"float32\"),\n",
" \"idio_vol\": np.full(N_ASSETS, SIGMA_IDIO, dtype=\"float32\"),\n",
" }\n",
")\n",
"\n",
"# Mcap drifts by a small lognormal walk over time so the upload isn't trivially flat.\n",
"mcap_walk = np.exp(\n",
" np.cumsum(rng.normal(0.0, 0.005, size=(T, N_ASSETS)), axis=0)\n",
").astype(\"float32\")\n",
"\n",
"# Cross-join dates × assets, then attach the time-varying market cap as a\n",
"# column. Polars cross-join order is `(date0, asset0), (date0, asset1), …,\n",
"# (date1, asset0), …`, which is exactly the C-order of `mcap_walk[t, i]`.\n",
"df_market_cap = (\n",
" pl.DataFrame({\"date\": dates})\n",
" .join(df_assets, how=\"cross\")\n",
" .with_columns(\n",
" pl.lit(\"ticker\").alias(\"asset_id_type\"),\n",
" pl.lit(\"USD\").alias(\"ccy\"),\n",
" (pl.col(\"base_mcap\") * pl.Series(mcap_walk.reshape(-1)))\n",
" .cast(pl.Float32)\n",
" .alias(\"market_cap\"),\n",
" )\n",
" .select(\n",
" \"date\", \"asset_id\", \"asset_id_type\", \"ccy\", \"market_cap\", \"volume\", \"idio_vol\"\n",
" )\n",
")\n",
"df_market_cap.head()"
]
},
{
"cell_type": "markdown",
"id": "14b9d092",
"metadata": {},
"source": [
"## Step 3 — `price` (the daily market record)\n",
"\n",
"`price` answers *what did each asset do today?* It carries close, daily\n",
"total return, and a delisted flag. Primary key is `(date, asset_id,\n",
"asset_id_type, ccy)` — `ccy` is part of the key so the same asset can\n",
"carry prices in multiple currencies simultaneously.\n",
"\n",
"We compound the simulated per-asset returns from a base of 100. There's\n",
"one wrinkle worth calling out: each asset's **first** row must have\n",
"`return = null` and serve as the price base. Without that base row the\n",
"engine has no `p[t-1]` against which to reconstruct the first realized\n",
"return. We use a single base row per asset at `dates[0]` because every\n",
"asset is alive on day zero in this simulation — a real pipeline with\n",
"new listings places the base row at each asset's first appearance.\n"
]
},
{
"cell_type": "code",
"execution_count": 7,
"id": "8b52f054",
"metadata": {
"lines_to_next_cell": 2
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"price rows: 7830 (= 30 base + 7800 observed)\n"
]
},
{
"data": {
"text/html": [
"
\n",
"shape: (5, 7)
date
asset_id
asset_id_type
ccy
close
return
delisted
date
str
str
str
f32
f32
bool
2024-01-02
"A0000"
"ticker"
"USD"
100.0
null
false
2024-01-02
"A0001"
"ticker"
"USD"
100.0
null
false
2024-01-02
"A0002"
"ticker"
"USD"
100.0
null
false
2024-01-02
"A0003"
"ticker"
"USD"
100.0
null
false
2024-01-02
"A0004"
"ticker"
"USD"
100.0
null
false
"
],
"text/plain": [
"shape: (5, 7)\n",
"┌────────────┬──────────┬───────────────┬─────┬───────┬────────┬──────────┐\n",
"│ date ┆ asset_id ┆ asset_id_type ┆ ccy ┆ close ┆ return ┆ delisted │\n",
"│ --- ┆ --- ┆ --- ┆ --- ┆ --- ┆ --- ┆ --- │\n",
"│ date ┆ str ┆ str ┆ str ┆ f32 ┆ f32 ┆ bool │\n",
"╞════════════╪══════════╪═══════════════╪═════╪═══════╪════════╪══════════╡\n",
"│ 2024-01-02 ┆ A0000 ┆ ticker ┆ USD ┆ 100.0 ┆ null ┆ false │\n",
"│ 2024-01-02 ┆ A0001 ┆ ticker ┆ USD ┆ 100.0 ┆ null ┆ false │\n",
"│ 2024-01-02 ┆ A0002 ┆ ticker ┆ USD ┆ 100.0 ┆ null ┆ false │\n",
"│ 2024-01-02 ┆ A0003 ┆ ticker ┆ USD ┆ 100.0 ┆ null ┆ false │\n",
"│ 2024-01-02 ┆ A0004 ┆ ticker ┆ USD ┆ 100.0 ┆ null ┆ false │\n",
"└────────────┴──────────┴───────────────┴─────┴───────┴────────┴──────────┘"
]
},
"execution_count": 7,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"# Compound returns to a price path starting at 100.\n",
"close = np.empty((T, N_ASSETS), dtype=\"float32\")\n",
"close[0] = 100.0\n",
"close[1:] = 100.0 * np.cumprod(1.0 + r[1:], axis=0)\n",
"\n",
"PRICE_COLS = [\"date\", \"asset_id\", \"asset_id_type\", \"ccy\", \"close\", \"return\", \"delisted\"]\n",
"\n",
"# Base rows: t=0 for every asset, close=100, return=null (no prior price).\n",
"df_price_base = pl.DataFrame({\"asset_id\": assets}).select(\n",
" pl.lit(dates[0]).alias(\"date\"),\n",
" \"asset_id\",\n",
" pl.lit(\"ticker\").alias(\"asset_id_type\"),\n",
" pl.lit(\"USD\").alias(\"ccy\"),\n",
" pl.lit(100.0, dtype=pl.Float32).alias(\"close\"),\n",
" pl.lit(None, dtype=pl.Float32).alias(\"return\"),\n",
" pl.lit(False).alias(\"delisted\"),\n",
")\n",
"\n",
"# Realized rows: t=1..T-1, close compounded, return = r[t]. Cross-join\n",
"# emits rows in (date, asset) order, matching the C-order reshape of\n",
"# close[1:] and r[1:].\n",
"df_price_obs = (\n",
" pl.DataFrame({\"date\": dates[1:]})\n",
" .join(pl.DataFrame({\"asset_id\": assets}), how=\"cross\")\n",
" .select(\n",
" \"date\",\n",
" \"asset_id\",\n",
" pl.lit(\"ticker\").alias(\"asset_id_type\"),\n",
" pl.lit(\"USD\").alias(\"ccy\"),\n",
" pl.Series(\"close\", close[1:].reshape(-1), dtype=pl.Float32),\n",
" pl.Series(\"return\", r[1:].reshape(-1), dtype=pl.Float32),\n",
" pl.lit(False).alias(\"delisted\"),\n",
" )\n",
")\n",
"\n",
"df_price = pl.concat([df_price_base, df_price_obs]).sort(\"date\", \"asset_id\")\n",
"print(\n",
" f\"price rows: {df_price.height} (= {N_ASSETS} base + {N_ASSETS * (T - 1)} observed)\"\n",
")\n",
"df_price.head()"
]
},
{
"cell_type": "markdown",
"id": "788afa8f",
"metadata": {},
"source": [
"## Step 4 — `exposures` (long format)\n",
"\n",
"The exposures upload is long-format with schema `(date, asset_id,\n",
"asset_id_type, factor_group, factor, exposure)`. Dense vs sparse routing\n",
"is decided by the engine at dataset-construction time based on which\n",
"factor groups you declare in `dense_factor_groups` vs\n",
"`sparse_factor_groups` on `RootRiskDatasetSettings` — there's no\n",
"per-row flag.\n",
"\n",
"Conceptually:\n",
"\n",
"* **Dense factor groups** (`market`, `style` here) carry a continuous\n",
" value per asset per day. The engine treats every asset as loading on\n",
" every factor in a dense group. We upload one row per asset-day-factor.\n",
"* **Sparse factor groups** (`industry`, `estimation_universe` here)\n",
" only carry the *nonzero* loadings — structural zeros are never\n",
" uploaded. The typical case is one-hot (a pure-play company has a\n",
" single row per day naming its industry), but multi-row spreads\n",
" are valid too: a conglomerate with 40% `TECH` and 60% `MATERIALS`\n",
" contributes two rows per day with fractional exposures. The\n",
" engine routes the sparse block through the path where\n",
" `thin_category_shrinkage` actually fires.\n",
"\n",
"We build the three long frames from the static profile and concatenate.\n"
]
},
{
"cell_type": "code",
"execution_count": 8,
"id": "5e89a085",
"metadata": {
"lines_to_next_cell": 2
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"exposures rows: 46980 (market=7830, style=23490, industry=7830, estu=7830)\n"
]
},
{
"data": {
"text/html": [
"
\n",
"shape: (5, 6)
date
asset_id
asset_id_type
factor_group
factor
exposure
date
str
str
str
str
f32
2024-01-02
"A0000"
"ticker"
"market"
"market"
1.0
2024-01-02
"A0001"
"ticker"
"market"
"market"
1.0
2024-01-02
"A0002"
"ticker"
"market"
"market"
1.0
2024-01-02
"A0003"
"ticker"
"market"
"market"
1.0
2024-01-02
"A0004"
"ticker"
"market"
"market"
1.0
"
],
"text/plain": [
"shape: (5, 6)\n",
"┌────────────┬──────────┬───────────────┬──────────────┬────────┬──────────┐\n",
"│ date ┆ asset_id ┆ asset_id_type ┆ factor_group ┆ factor ┆ exposure │\n",
"│ --- ┆ --- ┆ --- ┆ --- ┆ --- ┆ --- │\n",
"│ date ┆ str ┆ str ┆ str ┆ str ┆ f32 │\n",
"╞════════════╪══════════╪═══════════════╪══════════════╪════════╪══════════╡\n",
"│ 2024-01-02 ┆ A0000 ┆ ticker ┆ market ┆ market ┆ 1.0 │\n",
"│ 2024-01-02 ┆ A0001 ┆ ticker ┆ market ┆ market ┆ 1.0 │\n",
"│ 2024-01-02 ┆ A0002 ┆ ticker ┆ market ┆ market ┆ 1.0 │\n",
"│ 2024-01-02 ┆ A0003 ┆ ticker ┆ market ┆ market ┆ 1.0 │\n",
"│ 2024-01-02 ┆ A0004 ┆ ticker ┆ market ┆ market ┆ 1.0 │\n",
"└────────────┴──────────┴───────────────┴──────────────┴────────┴──────────┘"
]
},
"execution_count": 8,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"# Per-asset static frames used by the cross joins below. Keeping the\n",
"# static profile separate from the date axis is the canonical way to\n",
"# build long-format exposures in Polars.\n",
"df_asset_industry = pl.DataFrame({\"asset_id\": assets, \"industry\": list(asset_industry)})\n",
"df_asset_style = pl.DataFrame(\n",
" {\n",
" \"asset_id\": np.repeat(assets, len(STYLES)),\n",
" \"style\": STYLES * N_ASSETS,\n",
" \"loading\": asset_style_loadings.reshape(-1).astype(\"float32\"),\n",
" }\n",
")\n",
"df_dates = pl.DataFrame({\"date\": dates})\n",
"\n",
"# market: dense, every asset loads 1.0 (this is the intercept).\n",
"df_exp_market = df_dates.join(\n",
" pl.DataFrame({\"asset_id\": assets}), how=\"cross\"\n",
").with_columns(\n",
" pl.lit(\"ticker\").alias(\"asset_id_type\"),\n",
" pl.lit(\"market\").alias(\"factor_group\"),\n",
" pl.lit(\"market\").alias(\"factor\"),\n",
" pl.lit(1.0, dtype=pl.Float32).alias(\"exposure\"),\n",
")\n",
"\n",
"# style: dense, one row per (date, asset, style). Loadings are constant\n",
"# per asset in this simulation — production pipelines would update them\n",
"# daily.\n",
"df_exp_style = df_dates.join(df_asset_style, how=\"cross\").select(\n",
" \"date\",\n",
" \"asset_id\",\n",
" pl.lit(\"ticker\").alias(\"asset_id_type\"),\n",
" pl.lit(\"style\").alias(\"factor_group\"),\n",
" pl.col(\"style\").alias(\"factor\"),\n",
" pl.col(\"loading\").alias(\"exposure\"),\n",
")\n",
"\n",
"# industry: sparse / categorical. One row per asset-day naming the\n",
"# asset's industry.\n",
"df_exp_industry = df_dates.join(df_asset_industry, how=\"cross\").select(\n",
" \"date\",\n",
" \"asset_id\",\n",
" pl.lit(\"ticker\").alias(\"asset_id_type\"),\n",
" pl.lit(\"industry\").alias(\"factor_group\"),\n",
" pl.col(\"industry\").alias(\"factor\"),\n",
" pl.lit(1.0, dtype=pl.Float32).alias(\"exposure\"),\n",
")\n",
"\n",
"# estimation_universe: sparse, every asset is in our tutorial's estu.\n",
"# Production deployments use this group to declare which assets enter\n",
"# the regression on each day (e.g. \"top 3000 by mcap\").\n",
"df_exp_estu = df_dates.join(\n",
" pl.DataFrame({\"asset_id\": assets}), how=\"cross\"\n",
").with_columns(\n",
" pl.lit(\"ticker\").alias(\"asset_id_type\"),\n",
" pl.lit(\"estimation_universe\").alias(\"factor_group\"),\n",
" pl.lit(\"estimation_universe\").alias(\"factor\"),\n",
" pl.lit(1.0, dtype=pl.Float32).alias(\"exposure\"),\n",
")\n",
"\n",
"df_exposures = pl.concat([df_exp_market, df_exp_style, df_exp_industry, df_exp_estu])\n",
"print(\n",
" f\"exposures rows: {df_exposures.height} \"\n",
" f\"(market={df_exp_market.height}, style={df_exp_style.height}, \"\n",
" f\"industry={df_exp_industry.height}, estu={df_exp_estu.height})\"\n",
")\n",
"df_exposures.head()"
]
},
{
"cell_type": "markdown",
"id": "964284a7",
"metadata": {},
"source": [
"## Step 5 — `exchange_dates` (non-trading days)\n",
"\n",
"`exchange_dates` carries every **non-trading day** per exchange —\n",
"weekends *and* holidays. The engine doesn't infer weekends; if a\n",
"Saturday isn't in this upload it's treated as a trading day. The\n",
"trading calendar for an exchange is `(every date in the dataset) ∖\n",
"(rows in this upload for that exchange)`.\n",
"\n",
"We mark both — every Sat/Sun in 2024 plus the US federal holidays for\n",
"XNYS. A production pipeline would generate this from\n",
"`pandas.tseries.holiday` plus a weekday filter, or an\n",
"exchange-calendar service.\n"
]
},
{
"cell_type": "code",
"execution_count": 9,
"id": "4491bcc0",
"metadata": {
"lines_to_next_cell": 2
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"non-trading days: 113 (104 weekends + 9 holidays)\n"
]
},
{
"data": {
"text/html": [
"
\n",
"shape: (5, 2)
date
exchange
date
str
2024-01-06
"XNYS"
2024-01-07
"XNYS"
2024-01-13
"XNYS"
2024-01-14
"XNYS"
2024-01-15
"XNYS"
"
],
"text/plain": [
"shape: (5, 2)\n",
"┌────────────┬──────────┐\n",
"│ date ┆ exchange │\n",
"│ --- ┆ --- │\n",
"│ date ┆ str │\n",
"╞════════════╪══════════╡\n",
"│ 2024-01-06 ┆ XNYS │\n",
"│ 2024-01-07 ┆ XNYS │\n",
"│ 2024-01-13 ┆ XNYS │\n",
"│ 2024-01-14 ┆ XNYS │\n",
"│ 2024-01-15 ┆ XNYS │\n",
"└────────────┴──────────┘"
]
},
"execution_count": 9,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"us_holidays_2024 = {\n",
" dt.date(2024, 1, 1), # New Year\n",
" dt.date(2024, 1, 15), # MLK\n",
" dt.date(2024, 2, 19), # Presidents'\n",
" dt.date(2024, 3, 29), # Good Friday\n",
" dt.date(2024, 5, 27), # Memorial\n",
" dt.date(2024, 6, 19), # Juneteenth\n",
" dt.date(2024, 7, 4), # Independence\n",
" dt.date(2024, 9, 2), # Labor\n",
" dt.date(2024, 11, 28), # Thanksgiving\n",
" dt.date(2024, 12, 25), # Christmas\n",
"}\n",
"# Weekends + holidays = every non-trading day in range. New Year's Day\n",
"# (Jan 1) precedes the Jan-2 start, so only 9 of the 10 listed holidays\n",
"# fall in range — recompute the split from the rows so the parts sum to\n",
"# the total.\n",
"non_trading_days = [d for d in all_days if d.weekday() >= 5 or d in us_holidays_2024]\n",
"df_exchange_dates = pl.DataFrame(\n",
" {\n",
" \"date\": non_trading_days,\n",
" \"exchange\": [\"XNYS\"] * len(non_trading_days),\n",
" }\n",
")\n",
"n_weekend = sum(1 for d in non_trading_days if d.weekday() >= 5)\n",
"n_holiday = len(non_trading_days) - n_weekend\n",
"print(\n",
" f\"non-trading days: {len(non_trading_days)} \"\n",
" f\"({n_weekend} weekends + {n_holiday} holidays)\"\n",
")\n",
"df_exchange_dates.head()"
]
},
{
"cell_type": "markdown",
"id": "d09e8878",
"metadata": {},
"source": [
"## Step 6 — `exchange_rates` (USD-base FX)\n",
"\n",
"`exchange_rates` carries USD-base FX per `(date, ccy)`: **how many\n",
"units of `ccy` equal one USD on that date**. So JPY is a *large*\n",
"number (≈ 150) and EUR is a *small* one (≈ 0.93), because one USD\n",
"buys many yen but less than one euro. Equivalently: to convert a\n",
"foreign-currency price to USD you **divide** by `fx_rate`.\n",
"\n",
"We're USD-only in this tutorial, so every row has `fx_rate = 1.0`.\n",
"A multi-currency pipeline uploads one row per currency per date.\n"
]
},
{
"cell_type": "code",
"execution_count": 10,
"id": "d1764b5d",
"metadata": {
"lines_to_next_cell": 2
},
"outputs": [
{
"data": {
"text/html": [
"
\n",
"shape: (5, 3)
date
ccy
fx_rate
date
str
f32
2024-01-02
"USD"
1.0
2024-01-03
"USD"
1.0
2024-01-04
"USD"
1.0
2024-01-05
"USD"
1.0
2024-01-08
"USD"
1.0
"
],
"text/plain": [
"shape: (5, 3)\n",
"┌────────────┬─────┬─────────┐\n",
"│ date ┆ ccy ┆ fx_rate │\n",
"│ --- ┆ --- ┆ --- │\n",
"│ date ┆ str ┆ f32 │\n",
"╞════════════╪═════╪═════════╡\n",
"│ 2024-01-02 ┆ USD ┆ 1.0 │\n",
"│ 2024-01-03 ┆ USD ┆ 1.0 │\n",
"│ 2024-01-04 ┆ USD ┆ 1.0 │\n",
"│ 2024-01-05 ┆ USD ┆ 1.0 │\n",
"│ 2024-01-08 ┆ USD ┆ 1.0 │\n",
"└────────────┴─────┴─────────┘"
]
},
"execution_count": 10,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"df_exchange_rates = pl.DataFrame(\n",
" {\n",
" \"date\": dates,\n",
" \"ccy\": [\"USD\"] * T,\n",
" \"fx_rate\": np.ones(T, dtype=\"float32\"),\n",
" }\n",
")\n",
"df_exchange_rates.head()"
]
},
{
"cell_type": "markdown",
"id": "41e5f651",
"metadata": {
"lines_to_next_cell": 2
},
"source": [
"## Upload\n",
"\n",
"Each of the six frames goes to its own upload data type via the\n",
"Uploaders API. `create_or_replace_dataset(name)` creates the upload\n",
"dataset (or wipes and recreates if it already exists);\n",
"`fast_commit(df, mode=\"overwrite\")` stages and commits in one shot.\n",
"Every frame is built complete in-memory, so each upload is a single\n",
"`\"overwrite\"` commit.\n"
]
},
{
"cell_type": "code",
"execution_count": 11,
"id": "2f0e037c",
"metadata": {
"lines_to_next_cell": 2
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
" idmap → tutorial-idmap ( 30 rows)\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
" market_cap → tutorial-market-cap ( 7,830 rows)\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
" price → tutorial-price ( 7,830 rows)\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
" exposures → tutorial-exposures ( 46,980 rows)\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
" exchange_dates → tutorial-exchange-dates ( 113 rows)\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
" exchange_rates → tutorial-exchange-rates ( 261 rows)\n"
]
}
],
"source": [
"def upload(data_type: str, name: str, df: pl.DataFrame) -> None:\n",
" uploader = bln.equity.uploaders.get_data_type(data_type)\n",
" ds = uploader.create_or_replace_dataset(name)\n",
" ds.fast_commit(df, mode=\"overwrite\")\n",
" print(f\" {data_type:15s} → {name:28s} ({df.height:>7,} rows)\")\n",
"\n",
"\n",
"upload(\"idmap\", \"tutorial-idmap\", df_idmap)\n",
"upload(\"market_cap\", \"tutorial-market-cap\", df_market_cap)\n",
"upload(\"price\", \"tutorial-price\", df_price)\n",
"upload(\"exposures\", \"tutorial-exposures\", df_exposures)\n",
"upload(\"exchange_dates\", \"tutorial-exchange-dates\", df_exchange_dates)\n",
"upload(\"exchange_rates\", \"tutorial-exchange-rates\", df_exchange_rates)"
]
},
{
"cell_type": "markdown",
"id": "9cd693d9",
"metadata": {},
"source": [
"## Create the root dataset\n",
"\n",
"`RootRiskDatasetSettings` ties the six uploads into a single risk\n",
"dataset that the rest of the API targets.\n",
"\n",
"**Source references.** Each `*_source=` parameter is the name of an\n",
"upload dataset you created above. The engine resolves them lazily —\n",
"no data is actually read here.\n",
"\n",
"**`dense_factor_groups` vs `sparse_factor_groups`.** These two dicts\n",
"decide which regression block the engine routes each factor group\n",
"through:\n",
"\n",
"* **Dense** — continuous loading per asset per factor (`market`,\n",
" `style`). Every asset loads on every factor in the group.\n",
"* **Sparse** — only the *nonzero* loadings are uploaded\n",
" (`industry`, `estimation_universe`). One-hot membership is the\n",
" common case, but multi-row spreads are valid: a conglomerate with\n",
" partial loadings across several industries contributes one row\n",
" per nonzero entry. The sparse block is also where\n",
" `thin_category_shrinkage` fires when the model runs (Step 8).\n",
"\n",
"**The dict *value* — `None` vs a hierarchy upload source.** Each\n",
"entry pairs a factor-group name with an optional **factor\n",
"hierarchy**: a tree that organises the group's leaves into parent\n",
"buckets so reports can roll up (GICS sub-industry → industry →\n",
"sector, value sub-styles → value, etc.). The value is either:\n",
"\n",
"* the *name* of a hierarchy upload dataset you committed alongside\n",
" the other six (a separate recipe covers this), or\n",
"* `None`, which tells the engine to **synthesise a flat one-level\n",
" hierarchy** on the fly: each factor in the group becomes its own\n",
" leaf with no parent, no roll-up structure.\n",
"\n",
"`None` is the right choice for `market` and `estimation_universe`\n",
"(one factor each — nothing to roll up) and a perfectly fine\n",
"starting point for `style` and `industry` if you don't have a tree\n",
"yet. You can replace `None` with a hierarchy source later without\n",
"touching any of the six base uploads.\n",
"\n",
"**`as_of_date`** snapshots the `idmap` — only rows whose\n",
"`start_date <= as_of_date` are retained when the dataset's\n",
"`IdMapper` is built. Use the most recent date the idmap is\n",
"authoritative for.\n",
"\n",
"**`master_id_type` / `master_id_core_type`** must match the\n",
"`from_id_type` / `to_id_type` of the master-to-core projection rows\n",
"in your idmap upload (Step 1).\n"
]
},
{
"cell_type": "code",
"execution_count": 12,
"id": "a2c49bf8",
"metadata": {
"lines_to_next_cell": 2
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"id types: ['ticker', 'ticker_core']\n",
"exchanges: ['XNYS']\n",
"categorical hierarchies: ['industry', 'estimation_universe']\n",
"continuous hierarchies: ['market', 'style']\n"
]
}
],
"source": [
"settings = RootRiskDatasetSettings(\n",
" market_cap_source=\"tutorial-market-cap\",\n",
" price_source=\"tutorial-price\",\n",
" exposures_source=\"tutorial-exposures\",\n",
" exchange_dates_source=\"tutorial-exchange-dates\",\n",
" exchange_rates_source=\"tutorial-exchange-rates\",\n",
" idmap_source=\"tutorial-idmap\",\n",
" dense_factor_groups={\"market\": None, \"style\": None},\n",
" sparse_factor_groups={\"industry\": None, \"estimation_universe\": None},\n",
" master_id_type=\"ticker\",\n",
" master_id_core_type=\"ticker_core\",\n",
" as_of_date=dates[-1],\n",
")\n",
"\n",
"bln.equity.riskdatasets.delete_dataset_if_exists(\"tutorial-custom\")\n",
"dataset = bln.equity.riskdatasets.create_dataset(\"tutorial-custom\", settings)\n",
"props = dataset.describe()\n",
"print(\"id types: \", props.universe_settings_menu.id_types)\n",
"print(\n",
" \"exchanges: \",\n",
" props.universe_settings_menu.calendar_settings_menu.exchanges,\n",
")\n",
"print(\n",
" \"categorical hierarchies:\",\n",
" list(props.universe_settings_menu.categorical_hierarchies.keys()),\n",
")\n",
"print(\n",
" \"continuous hierarchies: \",\n",
" list(props.exposure_settings_menu.continuous_hierarchies.keys()),\n",
")"
]
},
{
"cell_type": "markdown",
"id": "de928876",
"metadata": {},
"source": [
"## Fit a factor model\n",
"\n",
"`FactorRiskModelSettings` declares one exposure group per factor block we\n",
"uploaded. We build the settings dataset-agnostic, then bind them to our\n",
"new dataset at load time via `.with_dataset(\"tutorial-custom\")` — the\n",
"same pattern the other tutorials use.\n",
"\n",
"`zero_sum_constraints={\"industry\": \"mcap_weighted\"}` matches the\n",
"constraint we applied in the ground-truth simulation, so the engine\n",
"estimates industry returns in the same gauge.\n",
"\n",
"`thin_category_shrinkage={\"industry\": 10.0}` shrinks any industry whose\n",
"mcap-weighted effective sample size falls below `N_min = 10` toward\n",
"zero — a guard against thin-industry blow-ups. With six industries and\n",
"30 assets we're nowhere near that threshold here, but setting it makes\n",
"the configuration realistic.\n"
]
},
{
"cell_type": "code",
"execution_count": 13,
"id": "67804260",
"metadata": {
"lines_to_next_cell": 2
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"fret shape: (252, 11)\n"
]
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"factors: {'industry': 6, 'market': 1, 'style': 3}\n"
]
},
{
"data": {
"text/html": [
"
\n",
"shape: (5, 11)
date
industry.CONSUMER
industry.ENERGY
industry.FINS
industry.HEALTH
industry.MATERIALS
industry.TECH
market.market
style.momentum
style.size
style.value
date
f32
f32
f32
f32
f32
f32
f32
f32
f32
f32
2024-01-02
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
0.0
2024-01-03
-0.001508
-0.000547
0.002913
-0.004646
0.00041
0.002329
0.007626
0.00291
-0.006499
0.004172
2024-01-04
0.005483
0.002196
0.002705
-0.010029
-0.003685
-0.000647
0.0086
0.002396
0.001811
-0.00352
2024-01-05
-0.005219
0.000634
0.000258
0.008118
0.001262
-0.000469
0.02151
-0.001198
0.016389
-0.007198
2024-01-08
-0.003841
-0.001597
0.00307
0.000216
0.000478
0.001505
0.025358
-0.009139
-0.0055
-0.002657
"
],
"text/plain": [
"shape: (5, 11)\n",
"┌───────────┬───────────┬───────────┬───────────┬───┬───────────┬───────────┬───────────┬──────────┐\n",
"│ date ┆ industry. ┆ industry. ┆ industry. ┆ … ┆ market.ma ┆ style.mom ┆ style.siz ┆ style.va │\n",
"│ --- ┆ CONSUMER ┆ ENERGY ┆ FINS ┆ ┆ rket ┆ entum ┆ e ┆ lue │\n",
"│ date ┆ --- ┆ --- ┆ --- ┆ ┆ --- ┆ --- ┆ --- ┆ --- │\n",
"│ ┆ f32 ┆ f32 ┆ f32 ┆ ┆ f32 ┆ f32 ┆ f32 ┆ f32 │\n",
"╞═══════════╪═══════════╪═══════════╪═══════════╪═══╪═══════════╪═══════════╪═══════════╪══════════╡\n",
"│ 2024-01-0 ┆ 0.0 ┆ 0.0 ┆ 0.0 ┆ … ┆ 0.0 ┆ 0.0 ┆ 0.0 ┆ 0.0 │\n",
"│ 2 ┆ ┆ ┆ ┆ ┆ ┆ ┆ ┆ │\n",
"│ 2024-01-0 ┆ -0.001508 ┆ -0.000547 ┆ 0.002913 ┆ … ┆ 0.007626 ┆ 0.00291 ┆ -0.006499 ┆ 0.004172 │\n",
"│ 3 ┆ ┆ ┆ ┆ ┆ ┆ ┆ ┆ │\n",
"│ 2024-01-0 ┆ 0.005483 ┆ 0.002196 ┆ 0.002705 ┆ … ┆ 0.0086 ┆ 0.002396 ┆ 0.001811 ┆ -0.00352 │\n",
"│ 4 ┆ ┆ ┆ ┆ ┆ ┆ ┆ ┆ │\n",
"│ 2024-01-0 ┆ -0.005219 ┆ 0.000634 ┆ 0.000258 ┆ … ┆ 0.02151 ┆ -0.001198 ┆ 0.016389 ┆ -0.00719 │\n",
"│ 5 ┆ ┆ ┆ ┆ ┆ ┆ ┆ ┆ 8 │\n",
"│ 2024-01-0 ┆ -0.003841 ┆ -0.001597 ┆ 0.00307 ┆ … ┆ 0.025358 ┆ -0.009139 ┆ -0.0055 ┆ -0.00265 │\n",
"│ 8 ┆ ┆ ┆ ┆ ┆ ┆ ┆ ┆ 7 │\n",
"└───────────┴───────────┴───────────┴───────────┴───┴───────────┴───────────┴───────────┴──────────┘"
]
},
"execution_count": 13,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"riskmodel_settings = FactorRiskModelSettings(\n",
" universe=UniverseSettings(id_type=\"ticker\"),\n",
" exposures=ExposureSettings(\n",
" exposures=[\n",
" ContinuousExposureGroupSettings(hierarchy=\"market\"),\n",
" CategoricalExposureGroupSettings(hierarchy=\"industry\"),\n",
" ContinuousExposureGroupSettings(hierarchy=\"style\"),\n",
" ]\n",
" ),\n",
" modelconstruction=ModelConstructionSettings(\n",
" estimation_universe=UniverseSettings(\n",
" id_type=\"ticker\",\n",
" categorical_filters=[\n",
" CategoricalFilterSettings(hierarchy=\"estimation_universe\")\n",
" ],\n",
" ),\n",
" zero_sum_constraints={\"industry\": \"mcap_weighted\"},\n",
" thin_category_shrinkage={\"industry\": 10.0},\n",
" return_clip_bounds=(None, None),\n",
" ),\n",
")\n",
"\n",
"model = bln.equity.riskmodels.load(\n",
" riskmodel_settings.with_dataset(\"tutorial-custom\")\n",
").get_model()\n",
"df_fret = model.fret()\n",
"print(\"fret shape:\", df_fret.shape)\n",
"print(\"factors: \", {k: len(v) for k, v in model.factors().items()})\n",
"df_fret.head()"
]
},
{
"cell_type": "markdown",
"id": "a065ad68",
"metadata": {},
"source": [
"## Sanity check: did the engine recover what we simulated?\n",
"\n",
"Because we generated the factor returns ourselves we know what `fret()`\n",
"should look like. Join the estimated returns against `df_truth` and plot\n",
"estimated vs true per factor. The points should hug a 45° line —\n",
"not exactly (idio noise plus regression-weighting choices add scatter),\n",
"but tightly enough that the dataset is clearly wired correctly.\n"
]
},
{
"cell_type": "code",
"execution_count": 14,
"id": "b4707586",
"metadata": {
"lines_to_next_cell": 2
},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"df_compare rows: 2510 (dates kept: 251)\n",
"market n= 251 corr=+0.951\n",
"industry n= 1506 corr=+0.581\n",
"style n= 753 corr=+0.666\n"
]
},
{
"data": {
"image/png": "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",
"text/plain": [
""
]
},
"metadata": {},
"output_type": "display_data"
}
],
"source": [
"import matplotlib.pyplot as plt\n",
"\n",
"# fret returns 0.0 on edge dates where the regression can't run (no prior\n",
"# price). Drop those rows so the scatter only contains real comparisons.\n",
"nonzero_dates = (\n",
" df_fret.unpivot(index=\"date\", variable_name=\"factor\", value_name=\"estimated\")\n",
" .group_by(\"date\")\n",
" .agg(pl.col(\"estimated\").abs().sum().alias(\"absum\"))\n",
" .filter(pl.col(\"absum\") > 0)\n",
" .select(\"date\")\n",
")\n",
"\n",
"df_compare = (\n",
" df_fret.unpivot(index=\"date\", variable_name=\"factor\", value_name=\"estimated\")\n",
" .join(nonzero_dates, on=\"date\", how=\"inner\")\n",
" .join(df_truth, on=(\"date\", \"factor\"), how=\"inner\")\n",
")\n",
"print(f\"df_compare rows: {df_compare.height} (dates kept: {nonzero_dates.height})\")\n",
"\n",
"groups = [\"market\", \"industry\", \"style\"]\n",
"fig, axes = plt.subplots(1, len(groups), figsize=(13, 4))\n",
"for ax, g in zip(axes, groups):\n",
" sub = df_compare.filter(pl.col(\"factor\").str.starts_with(g + \".\"))\n",
" x = sub[\"true_return\"].to_numpy()\n",
" y = sub[\"estimated\"].to_numpy()\n",
" ax.scatter(x, y, s=8, alpha=0.5)\n",
" if x.size:\n",
" lo = float(min(x.min(), y.min()))\n",
" hi = float(max(x.max(), y.max()))\n",
" ax.plot([lo, hi], [lo, hi], color=\"black\", linewidth=0.5)\n",
" corr = float(np.corrcoef(x, y)[0, 1])\n",
" ax.set_title(f\"{g} (n={x.size}, corr={corr:.3f})\")\n",
" print(f\"{g:8s} n={x.size:5d} corr={corr:+.3f}\")\n",
" else:\n",
" ax.set_title(f\"{g} (no rows)\")\n",
" ax.set_xlabel(\"true\")\n",
" ax.set_ylabel(\"estimated\")\n",
"fig.tight_layout()\n",
"plt.show()"
]
},
{
"cell_type": "markdown",
"id": "cb308e8e",
"metadata": {},
"source": [
"## Out of scope (and where to look next)\n",
"\n",
"A few capabilities of `RootRiskDatasetSettings` we deliberately skipped\n",
"to keep this tutorial focused:\n",
"\n",
"* **`series_source`** — optional seventh upload of scalar time series\n",
" (risk-free rate, region-aggregate returns, etc.). Pass\n",
" `series_source=\"...\"` on `RootRiskDatasetSettings` after uploading a\n",
" frame with `(date, series, value)`.\n",
"* **Uploaded hierarchies** — both `dense_factor_groups` and\n",
" `sparse_factor_groups` accept an upload source name in place of\n",
" `None` to load a custom factor hierarchy (e.g. sector → industry →\n",
" sub-industry trees). The default `None` gives a flat one-level\n",
" hierarchy, which is what we used.\n",
"* **Catch-up / incremental uploads** — see\n",
" [Exposure Catch-up Upload](recipe_exposure_catch_up_upload.ipynb) for\n",
" the pattern of appending new dates to an existing upload dataset.\n",
"* **Building a derived dataset on top of this one** — once\n",
" `tutorial-custom` exists, you can pass `reference_dataset=\n",
" \"tutorial-custom\"` on `DerivedRiskDatasetSettings` to layer additional\n",
" exposures / filters / hierarchies. See\n",
" [Model Onboarding](recipe_model_onboarding.ipynb) and\n",
" [Risk Datasets](tutorial_datasets.ipynb).\n"
]
}
],
"metadata": {
"jupytext": {
"text_representation": {
"extension": ".py",
"format_name": "percent",
"format_version": "1.3",
"jupytext_version": "1.19.3"
}
},
"kernelspec": {
"display_name": ".venv",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.11.15"
}
},
"nbformat": 4,
"nbformat_minor": 5
}