Blame for README.md

---
license: other
language:
- en
pretty_name: Dragon Finance - rates
tags:
- finance
- financial-data
- us-financial-data
- rates
---
<style>
/* Document-grade card styling: structural rules only, no decorative treatment.
   Two deliberate constraints, both learned the hard way on this account:
   (1) no background is set anywhere, so the card inherits the Hub's own text colour and stays
       legible in both themes. The previous skin forced a dark ground and a neon palette on every
       heading, rule and table; HF strips @media from card <style> blocks, so a forced palette
       could never adapt to light mode.
   (2) rules use a translucent grey rather than a fixed colour, so a hairlines sits correctly on
       either background. Nothing here computes or restates a figure. */
body { font-family: -apple-system, BlinkMacSystemFont, "Segoe UI", Roboto, "Helvetica Neue", Arial, sans-serif;
       font-size: 15px; line-height: 1.62; }
h1 { font-size: 1.5em; font-weight: 650; letter-spacing: -.011em; margin: 0 0 .35em;
     padding-bottom: .35em; border-bottom: 2px solid #3d7ea6; }
h2 { font-size: 1.14em; font-weight: 650; margin: 2em 0 .6em; padding-bottom: .28em;
     border-bottom: 1px solid rgba(128,128,128,.38); }
h3 { font-size: 1.02em; font-weight: 650; margin: 1.5em 0 .5em; }
h4, h5, h6 { font-size: .9em; font-weight: 650; text-transform: uppercase; letter-spacing: .07em;
             margin: 1.3em 0 .4em; }
p, li { margin: .6em 0; }
strong, b { font-weight: 650; }
code { font-family: ui-monospace, SFMono-Regular, Menlo, Consolas, monospace; font-size: .9em;
       padding: .08em .34em; border: 1px solid rgba(128,128,128,.38); border-radius: 3px; }
pre { border: 1px solid rgba(128,128,128,.38); border-radius: 4px; padding: 12px 14px;
      overflow-x: auto; }
pre code { border: 0; padding: 0; font-size: .88em; }
blockquote { margin: 1em 0; padding: .15em 0 .15em 1em; border-left: 3px solid #3d7ea6; }
table { border-collapse: collapse; width: 100%; margin: .9em 0 1.4em; font-size: .93em; }
th { text-align: left; font-weight: 650; font-size: .8em; text-transform: uppercase;
     letter-spacing: .07em; padding: 8px 10px; border-bottom: 2px solid rgba(128,128,128,.38);
     white-space: nowrap; }
td { padding: 7px 10px; border-bottom: 1px solid rgba(128,128,128,.28); vertical-align: top; }
td:not(:first-child), th:not(:first-child) { font-variant-numeric: tabular-nums; }
a { text-decoration: none; border-bottom: 1px solid #3d7ea6; }
a:hover { border-bottom-width: 2px; }
hr { border: 0; border-top: 1px solid rgba(128,128,128,.38); margin: 1.7em 0; }
img { max-width: 100%; }
</style>



# Dragon Finance — rates

**stored:** 65.62 MB  ·  **shards:** 21  ·  **class:** rates  ·  **sources:** US + global official

US sovereign rates and public finance — the risk-free curve and federal cash flows.

Part of the **Dragon Finance Corpus** — public financial data from US federal agencies, world
official institutions and public registries. Nothing licensed, paywalled or registration-only.

Front door: [`Dragon-Finance-Corpus`](https://huggingface.co/datasets/dragonlimited/Dragon-Finance-Corpus)
— every card, the chunked corpus and the derived artefacts, all measured from the hub;
[collection](https://huggingface.co/collections/dragonlimited/dragon-finance-corpus-6ae5fbf5e8d41b80f0ca44f0)
— this repo and its siblings.



## Coverage

| | |
|---|---|
| Yield curve | Treasury daily yield curve, all maturities |
| Auctions | auction results via the Treasury fiscal API (auctions_query) |
| Fiscal | Treasury fiscal data: debt to the penny, average interest rates, interest expense |
| Money-market rates | NY Fed reference rates and SOMA holdings |
| Eurozone | ECB yield-curve and bank interest-rate flows |

## Size

| metric | value |
|---|---|
| stored | **65.62 MB** |
| shards | 21 |
| **training-ready tokens** | **0.4 M** (0.01% of the chunked corpus) |

## Training-ready text

Measured 2026-11-04 20:37 UTC over 48 of 48 shard(s). Tokens use the one convention these cards share - `chars / 4` over the chunked text - so each figure is reconcilable against the shards.

| source | chunks | text | tokens | share |
|---|---:|---:|---:|---:|
| `treasury` | 1,500 | 0.001 GB | 0.4 M | 0.00% |
| `treasury_fiscal` | 108 | 0.000 GB | 0.0 M | 0.00% |
| `bankofengland` | 6 | 0.000 GB | 0.0 M | 0.00% |

Class attribution uses the shard path's first segment, so a tier that shares a segment with
another class (`census`, `federalreserve`, `sec`) is attributed to its dominant class.

## Sources present

| source | jurisdiction | data | licence |
|---|---|---|---|
| bankofengland | n/a | n/a | n/a |
| US Treasury | US | daily yield curve | public domain |
| US Treasury Fiscal | US | average interest rates | public domain |
| ECB | EU | official euro reference rates | ECB terms |
| FRED (St. Louis Fed) | US | rates + macro series | CC-BY-4.0 |
| SEC municipal advisers | US | registered municipal advisor registrations | US public record |
| NY Fed | US | reference rates (EFFR/OBFR/SOFR) + SOMA holdings | public |
| progress.png | n/a | n/a | n/a |
| US Treasury | US | daily yield curve | public domain |

## What you can do with it

* Build discount curves
* Track issuance and auction demand
* Monitor federal interest cost

## For engineers

Data is stored as **tar shards** (the hub caps a directory at 10,000 files and these tiers
run to millions). Members keep their original names and relative paths.

```python
import tarfile
with tarfile.open("0000.tar") as tf:
    print(tf.getnames()[:5])
    tf.extractall("out/")
```

| shard | size |
|---|---:|
| `README.md` | 6.8 KB |
| `bankofengland/0000.tar` | 133.1 KB |
| `data/treasury/0000.tar` | 675.8 KB |
| `data/treasury_fiscal/0000.tar` | 174.1 KB |
| `ecb/flows/0000.tar` | 3.38 MB |
| `fred/rates/0000.tar` | 4.63 MB |

_Showing 6 of 21 shards._

Every shard records its **sha256** in the pipeline ledgers; shards are append-only, so a
resumed run re-uploads only what changed. A shard with a failed member is never marked
complete, so missed files are retried rather than dropped.

---

> **Not advice, no warranty, no endorsement.** Compiled from third-party public sources and
> provided *'as is'*: derived and model-generated content may be wrong, and source material is
> routinely restated, amended or superseded. No source agency, exchange or company has reviewed,
> approved or endorsed this dataset, and none is affiliated with its publisher. Do not use it to
> re-identify any person, to manipulate any market, or as the sole basis for any decision about
> any person, entity or instrument. **Terms of use:**
> [`DISCLAIMER.md`](https://huggingface.co/datasets/dragonlimited/Dragon-Finance-Corpus/blob/main/DISCLAIMER.md).